Rapid advancements in the understanding and manipulation of tumor-immune interactions have led to the approval of immune therapies for patients with non–small cell lung cancer. Certain immune checkpoint inhibitor therapies require the use of companion diagnostics, but methodologic variability has led to uncertainty around test selection and implementation in practice.
To develop evidence-based guideline recommendations for the testing of immunotherapy/immunomodulatory biomarkers, including programmed death ligand-1 (PD-L1) and tumor mutation burden (TMB), in patients with lung cancer.
The College of American Pathologists convened a panel of experts in non–small cell lung cancer and biomarker testing to develop evidence-based recommendations in accordance with the standards for trustworthy clinical practice guidelines established by the National Academy of Medicine. A systematic literature review was conducted to address 8 key questions. Using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach, recommendations were created from the available evidence, certainty of that evidence, and key judgments as defined in the GRADE Evidence to Decision framework.
Six recommendation statements were developed.
This guideline summarizes the current understanding and hurdles associated with the use of PD-L1 expression and TMB testing for immune checkpoint inhibitor therapy selection in patients with advanced non–small cell lung cancer and presents evidence-based recommendations for PD-L1 and TMB testing in the clinical setting.
The advent of immune modulatory therapy has led to a radical shift in the treatment paradigm for patients with a wide range of cancers. The clinical impact of these therapies, especially those targeting immune “checkpoints,” has been particularly profound for patients with non–small cell lung carcinoma (NSCLC). Clinical trials have demonstrated that drugs that block programmed death receptor-1 (PD-1, encoded by PDCD1) and programmed death ligand-1 (PD-L1, also known as B7H1, encoded by CD274) lead to significant improvements in both response and survival relative to conventional cytotoxic chemotherapy for patients with advanced-stage NSCLC. These trials have identified PD-L1 protein expression, based on immunohistochemistry (IHC), as a biomarker for improved benefit following anti–PD-1/PD-L1 therapies and combination therapies (hereafter referred to as immune checkpoint inhibitors, or ICIs). The biomarker testing space has been complicated, however, by the proliferation of PD-L1 assays and scoring criteria that have evolved with individual therapies and often for different tumor types, some of which have received companion diagnostic (CDx) approvals by regulatory agencies such as the US Food and Drug Administration (FDA) and the Health Products and Food Branch of Health Canada.1 At the same time, for reasons of cost and access, PD-L1 IHC antibodies and assays developed outside of the scope of randomized controlled trials (RCTs) have garnered widespread use,2 thus leading to confusion on the part of pathologists and clinicians about the best approach to biomarker testing to select patients for ICI therapy. In contrast to most genomic biomarkers (eg, anaplastic lymphoma kinase [ALK], epidermal growth factor receptor [EGFR]), which tend to represent relatively stable and binary data points in a patient’s tumor profile in treatment-naïve patients, PD-L1 expression is dynamic and heterogeneous, complicating the choice of sample for testing.3–5
At the same time, there is ongoing interest in genomic biomarkers in immunotherapy—in particular, tumor mutation burden (TMB)—that may be used in conjunction with or independent of PD-L1 status. Indeed, high TMB has been approved by the FDA as a cancer-type agnostic biomarker for the anti–PD-1 monoclonal antibody pembrolizumab, thus elevating this test to clinical relevance, despite numerous challenges relating to access, technical reproducibility, and biological relevance of the proposed cutoff.
The guideline’s primary goal is to develop evidence-based recommendations for the testing of immunotherapy biomarkers, including PD-L1 and TMB, in patients with NSCLC. Several ICI-based therapies have been approved by regulatory agencies globally in the first and second lines of therapy for patients with NSCLC. Companion diagnostics are required for certain therapies, but for reasons of cost and access to necessary reagents and equipment, variable methodology exists in practice. As a result, questions regarding assay interchangeability persist.
METHODS
This evidence-based guideline was developed following the standards by the National Academy of Medicine.6 A detailed description of the panel composition, conflict of interest (COI) policy, and systematic review methods used to create this guideline can be found in the online Evidence-Based Guidelines Development Methodology Manual (Methodology Manual).7
Guideline Panel
The College of American Pathologists (CAP), in collaboration with the American Society of Clinical Oncology (ASCO), Association for Molecular Pathology (AMP), International Association for the Study of Lung Cancer (IASLC), Pulmonary Pathology Society (PPS), and the patient advocacy organization LUNGevity Foundation, convened a multidisciplinary expert and advisory panel to develop the guideline and approved the appointment of the members. Members included practicing pathologists, biomedical scientists, oncologists, patient advocates, and a guideline methodologist. Panel members were selected to represent diverse laboratory environments and geographic locations to assure that multiple perspectives would be represented. The roles of each panel member are described in the Methodology Manual. Detailed information about the panel composition can be found in the supplemental digital content (SDC) at https://meridian.allenpress.com/aplm in the July 2024 table of contents. The expert panel (EP) met via teleconference and 1 in-person meeting, using a modified Delphi method to come to agreements about the guideline scope and recommendations. Work was also conducted via email communication.
Conflict of Interest
In accordance with the CAP COI policy, members of the EP disclosed all financial interests from 2 years prior to appointment through the development of the guideline. Complete disclosures of the EP members are listed in the Appendix. A detailed description of the policy is included in the Methodology Manual.
The majority of the EP (10 of 12 members) was assessed as having no relevant COI. Disclosures of interest judged by the oversight group to be manageable conflicts are as follows. Mark Awad: consulting fees or advisory board, Affini-T Therapeutics, Inc (Watertown, Massachusetts), AstraZeneca (Wilmington, Delaware), Bristol-Myers Squibb (Princeton, New Jersey), EMD Serono (Rockland, Massachusetts), Genentech (South San Francisco, California), Gritstone bio (Emeryville, California), Foundation Medicine (Boston, Massachusetts), Instil Bio (Dallas, Texas), Janssen Oncology (Titusville, New Jersey), Merck (Rahway, New Jersey), Mirati Therapeutics Inc (San Diego, California), Novartis (East Hanover, New Jersey), Pfizer (New York, New York), Regeneron Pharmaceuticals (Tarrytown, New York); research grants, AstraZeneca (Wilmington, Delaware), Bristol-Myers Squibb (Princeton, New Jersey). Lauren Ritterhouse: employment and stock options, Foundation Medicine (Boston, Massachusetts).
The CAP provided funding for the administration of the project; no industry funds were used in the development of the guideline. All panel members volunteered their time and were not compensated for their involvement, except for the contracted methodologist.
Guideline Objectives
The panel addressed the following overarching questions, “Does PD-1/PD-L1 status and TMB improve clinical outcomes in patients with NSCLC who are being considered for ICI therapy?” and “What testing and specimen requirements provide accurate test results for PD-1/PD-L1 and TMB?” This led to the following key questions:
In patients with advanced-stage NSCLC who are being considered for ICI therapy, does PD-L1 and TMB testing improve treatment response rates and survival rates?
When selecting patients for anti–PD-1 and anti–PD-L1 therapy, does testing of different specimen types provide concordant clinical outcomes?
Does the use of ICI therapy in patients with advanced NSCLC with targetable ALK, EGFR, ROS proto-oncogene 1 (ROS1), or B-Raf proto-oncogene (BRAF) molecular alterations affect their long-term clinical outcomes?
When selecting patients for anti–PD-1 and anti–PD-L1 therapy, does TMB testing have the analytic validity to identify a complementary population who will benefit from therapy?
In patients with NSCLC with more than 1 available sample, do multiple samples provide concordant PD-L1 and TMB testing results and downstream clinical outcomes?
Does clinical validity of PD-L1 testing differ by levels of PD-L1 expression in tumor or immune cells?
How reproducible are PD-L1 tumor cell scores and immune cell scores across specimen types?
Do the available PD-L1 assays provide concordant expression profiles when evaluating the same sample, and which IHC expression cutoff provides the most reproducible expression categorization across the assays?
See Supplemental Table 1 in the SDC for more details.
Literature Search and Collection
A comprehensive literature search for relevant evidence was completed by a medical librarian. The search strategy was first constructed in Ovid MEDLINE (Wolters Kluwer Health, Inc) using controlled vocabulary and keywords agreed upon by the EP to reflect the population, intervention, comparator, and outcome elements, then translated into Embase (Elsevier) and Cochrane Library (John Wiley & Sons, Inc). Major concepts included NSCLC, PD-L1/PD-1, TMB, and laboratory testing. Limits were set to reflect the protocol inclusion/exclusion criteria, including (1) the publication date range of January 1, 2010, through the date each search was run; (2) language filters to capture only full-text articles available in English owing to time and financial constraints; and (3) publication limits to exclude letters, editorials, commentaries, and case studies. The Cochrane filter was applied to exclude animal studies. The database literature searches were initially run on October 16, 2019, and rerun on April 7, 2021, and May 13, 2022, to capture studies published since the initial searches were run. Supplemental searches to complement the database references were completed, and EP members were polled for relevant unpublished data at the onset of the project. See the SDC Supplemental Figures 1 and 2 for more information, including specific search strategies, supplemental search sources used, dates of search activity, and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) table that details the systematic review process.
Inclusion and Exclusion Criteria
Studies were selected for inclusion in the systematic review of evidence if they met the following criteria: (1) peer-reviewed articles published since January 1, 2010; (2) study population consisted of adult patients with early- or advanced-stage NSCLC either receiving or undergoing selection for checkpoint inhibitor therapy; (3) study compared, prospectively or retrospectively, laboratory testing methodologies for PD-L1; and (4) study included measurable data such as diagnostic test characteristics, accuracy of PD-L1 expression, survival outcomes, or treatment response.
Articles were excluded from the systematic review if (1) they were published before 2010; (2) they were editorials, letters, commentaries, and invited opinions; (3) the full-text article was not available in English; and (4) they did not address at least 1 key question or the outcomes of interest.
Assessing the Strength of Recommendations
Development of recommendations required the EP to review the identified evidence and make a series of key judgments, using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.8 See Supplemental Table 2 for the definitions of strength of recommendation. Supplemental Table 3 in the SDC provides a summary of the key judgments the panel considered, including the benefits and harms of each guideline statement using the GRADE Evidence to Decision (EtD) framework.9
Assessing Quality and Risk of Bias
Each study received a risk of bias assessment, and each recommendation was assigned an aggregate assessment of the certainty of evidence (Table 1). Refer to the SDC for definitions of the certainty of evidence (Supplemental Table 4) and for the individual study and aggregate risk of bias assessment (Supplemental Tables 5 through 9).
RESULTS
A total of 3089 studies met the eligibility requirements for screening. Based on review of these titles and abstracts, 356 articles met the inclusion criteria and continued to full-text review. A total of 121 articles were included for data extraction and qualitative analysis. Excluded articles were available as discussion or background references. Additional information is available in the SDC, including a PRISMA table outlining the details of the systematic review. Refer to the write-up for each recommendation for specific details about supporting evidence.
The EP convened by teleconferences and 1 face-to-face meeting to develop the scope, draft recommendations, review and respond to solicited feedback, and assess the certainty of evidence that supports the final recommendations presented herein. A modified Delphi technique was used for consensus decision-making to encourage balanced input and participation among members. An open comment period was posted on the CAP website (www.cap.org) from March 31 to April 23, 2021, during which the draft recommendation statements were available for public feedback. Refer to the SDC for more details.
The EP approved the final recommendations with a supermajority vote (ie, at least 75% of panel members in agreement). An independent review panel, masked to the EP and vetted through the COI process, recommended approval by the CAP Council on Scientific Affairs. The manuscript was also reviewed and approved by the collaborating associations. The final recommendations are summarized in Table 2.
Recommendation Statements
1. In patients with advanced non–small cell lung cancer, pathologists should use a validated PD-L1 immunohistochemistry expression assay, in conjunction with other targetable genomic biomarker assays where appropriate, to optimize selection for treatment with immune checkpoint inhibitors.
(Strength of Recommendation: Strong; Certainty of Evidence: Moderate)
The evidence for this statement included a total of 34 studies10–43 that evaluated overall survival (OS) rates and response rate (RR) following treatment with various immunotherapy agents in tumors with known PD-L1 expression status. The certainty of evidence was moderate for both outcomes of interest. From the available evidence, EP members concluded that OS and RR with immunotherapy were correlated with PD-L1 expression status. After discussions, EP members defined the benefits of PD-L1 expression detection using a validated IHC assay as moderate and the harms of this testing as small, and concluded that the benefits thus outweighed the harms. It is expected that this guidance will be acceptable to key stakeholders and feasible to implement. Refer to Supplemental Tables 5 through 8 for a summary of the risk of bias assessment for all included studies and the certainty of evidence assessment for all outcomes informing the statement. Supplemental Table 3 summarizes the EtD framework.
Published studies have demonstrated a statistically significant correlation between the presence and extent of PD-L1 expression in tumor tissue samples or tumor proportion score (TPS) and patient response and survival following immunotherapy with PD-1 and PD-L1 inhibitors given alone or in combination with chemotherapy and/or cytotoxic T-lymphocyte–associated protein-4 (CTLA-4) inhibitors (Table 3). To date, these associations have been most widely demonstrated in the subset of patients with stage IV metastatic nonsquamous (mainly adenocarcinoma) and squamous cell NSCLC. PD-L1 ICI therapy with atezolizumab has also shown a disease-free survival benefit in the adjuvant setting for patients with PD-L1–positive (TPS ≥1%), resected NSCLC.44 Similarly, PD-1 ICI therapy with pembrolizumab has been approved in the adjuvant setting for patients with resected IB-IIIA NSCLC, regardless of PD-L1 expression status. PD-1 therapy with nivolumab in combination with chemotherapy has also been approved in the neoadjuvant setting for patients with surgically resectable NSCLC irrespective of PD-L1 status.45 However, owing to the timing of the literature review and guideline drafting (which occurred primarily before the trials supporting these approvals were published), data related to the use of PD-L1 and related biomarkers in early-stage NSCLC are out of scope for this guideline.
Reported Overall Survival and Response Rates for First-Line Immune Checkpoint Inhibitors in Identified Randomized Controlled Trials

Although there is an association between PD-L1 expression and response for all lines of therapy in patients with advanced disease, these associations are strongest in the first-line setting. From phase-3 RCTs comparing PD-1/PD-L1 inhibitors to chemotherapy as first-line treatment in patients with advanced NSCLC, the FDA approved monotherapy with pembrolizumab, atezolizumab, or cemiplimab in patients whose tumors showed a PD-L1 TPS of 50% or more, and combination nivolumab + ipilimumab in patients with tumors expressing PD-L1 at TPS of 1% or more.9,45 Of note, the first-line atezolizumab approval also included patients with tumors harboring PD-L1–positive tumor-infiltrating immune cells covering 10% or more of the tumor area. A subsequent phase-3 RCT in patients with advanced NSCLC comparing first-line pembrolizumab monotherapy to chemotherapy showed a statistically significant survival benefit with pembrolizumab in the subgroup of patients with a PD-L1 TPS of 1% or more; however, this benefit was largely driven by superior outcomes in those with a TPS of 50% or more.46 Nevertheless, the FDA expanded the indication for first-line pembrolizumab monotherapy to include people with tumors with a PD-L1 TPS of 1% or more. However, this strategy has not been embraced globally, with other international regulatory agencies adhering to the threshold of 50% or more.47
Around the world, several other immunotherapy and chemoimmunotherapy combinations have also been approved in the first-line setting for patients with advanced NSCLC, based on clinical trials that have used a specific PD-L1 IHC assay with a scoring cut point to define PD-L1 positivity (Table 3). Some of these trials required certain levels of PD-L1 TPS for enrollment, but many did not, and while the benefits of chemotherapy-ICI combinations have been documented in patients across the PD-L1 TPS spectrum, most studies show improved outcomes with higher levels of PD-L1 expression. The EP recognizes that PD-L1 expression is not an absolute predictive biomarker, with responses and benefit seen in some patients with no visible expression in their tumor sample and lack of benefit in some patients despite high PD-L1 TPS. Nevertheless, considering the plethora of treatment options now available for patients with advanced NSCLC, PD-L1 TPS can be a useful biomarker to inform treatment decision-making and balance potential benefits and risks for individual patients.
Multiple clinical trials of ICI therapy for NSCLC have demonstrated a relative lack of benefit for those patients with driver alterations in the EGFR or ALK genes.12 Subsequently, many, but not all, trials of ICI or chemotherapy-ICI in NSCLC have excluded patients with EGFR and ALK alterations. Furthermore, prospective studies of treatment with concurrent durvalumab plus the EGFR tyrosine kinase inhibitor (TKI) osimertinib in patients with EGFR-mutant NSCLC were halted owing to unacceptable levels of interstitial pneumonitis (IP). An analysis of the FDA Adverse Event Reporting System found that the combination of nivolumab plus an EGFR TKI led to a significantly increased risk for development of IP relative to EGFR TKI therapy alone, with the IP events largely restricted to the Japanese population.48 However, a US-based single-center retrospective analysis also found that 15% of patients who received pembrolizumab followed by osimertinib had severe immune-related adverse events.49 In contrast to low response rates for immunotherapy in patients with EGFR and ALK alterations, patients with tumors harboring BRAF mutations appear to respond just as well to immunotherapy as those with wild-type tumors, whereas data on outcomes of immunotherapy in patients with other targetable driver alterations (eg, ROS1, RET, MET) are relatively limited and inconclusive.50–52 Patients with targetable oncogene-driven NSCLC derive substantial benefit from appropriate targeted therapy in the first-line setting,53–55 arguing for the value of obtaining comprehensive genomic profiling at the time of diagnosis with advanced disease.56
During the public comment period, of a total of 79 responders, 73 (92.4%) agreed or agreed with suggested modifications to the draft statement, 4 (5.06%) disagreed, and 2 (2.53%) were neutral. There were 17 written comments, many of which suggested that the final recommendation statement does not include the word advanced. The general comments also suggested that the FDA approval was for early stage, although the evidence included in the data review was on patients with late-stage NSCLC. Public comments also suggested that a discussion of the utility of PD-L1 in addition to a complete lung molecular biomarker profile is important. These comments were taken into consideration. While the recommendation remained the same, the comments were addressed in the discussion above.
2. Pathologists should ensure appropriate validation has been performed on all specimen types and fixatives.
Note: Specific validation requirements are out of scope with this guideline, and laboratories should refer to the Principles of Analytic Validation of Immunohistochemical Assays Guideline57 for details on how to validate IHC specimens.
(Strength of Recommendation: Conditional; Certainty of Evidence: Low)
The evidence base informing this statement is composed of 35 studies reporting on immunotherapy RR and survival rates,22,58 PD-L1 status concordance (Figure),59–85 diagnostic test characteristics of PD-L1 expression detection,71,77,82–84 and PD-L1 status using various specimen types.86,87 Both interobserver and intraobserver agreement for PD-L1 status using tissue69,75,88–90 and cytology specimens69,71,75,76,91 were also considered. The certainty of evidence across the 19 outcomes ranged from very low through moderate. EP members defined the benefits of PD-L1 expression detection using the best specimen available as moderate; however, the harms were also defined as moderate and the overall certainty of evidence was low, leading to the conclusion that balance of effects probably favored testing in the best available specimen. EP members also discussed that there was possibly important variability in the values and preferences of key stakeholders, but the guidance is expected to be acceptable and feasible to implement. Refer to Supplemental Tables 5 through 8 for a summary of the risk of bias assessment for all included studies and the certainty of evidence assessment for all outcomes informing the statement. Supplemental Table 3 summarizes the EtD framework.
PD-L1 tumor proportion score in cytology specimens versus histology sections. Reference standard defined as surgical resection for all studies except those denoted with an asterisk. In asterisk-denoted studies, the reference standard was a mixed FFPE sample of cell blocks, small biopsy samples, and surgical sections. Abbreviations: FFPE, formalin-fixed, paraffin-embedded; IV, inverse variance; PD-L1, programmed death ligand-1; RE, random effects.
PD-L1 tumor proportion score in cytology specimens versus histology sections. Reference standard defined as surgical resection for all studies except those denoted with an asterisk. In asterisk-denoted studies, the reference standard was a mixed FFPE sample of cell blocks, small biopsy samples, and surgical sections. Abbreviations: FFPE, formalin-fixed, paraffin-embedded; IV, inverse variance; PD-L1, programmed death ligand-1; RE, random effects.
Several variables may affect a given specimen’s suitability for testing, including age, site of origin, size (biopsy versus resection), and modality (histologic versus cytologic preparation). While several studies examining PD-L1 score concordance in relation to these variables were reviewed, only a handful analyzed clinical outcomes and response to ICI therapy. Based on the available evidence, it is difficult to determine what constitutes an optimal specimen for testing in patients with advanced NSCLC. Given that acquisition of material for testing entails a degree of morbidity depending on the patient’s condition and distribution of disease, the recommendation provides room for clinical judgment on the part of the pathologist and treating oncologist about selection of the most appropriate specimen, particularly when multiple specimens are available for testing.
The minimum criteria for adequacy as defined by the assay description for the PD-L1 IHC 22C3 Dako/Agilent pharmDx assay (Agilent Technologies, Inc, Santa Clara, California) is 100 viable cells.92 This minimum threshold is supported by 2 studies that demonstrate higher rates of positivity at both the 1% and 50% TPS cutoffs93,94 in cases where at least 100 viable cells are available for review. In specimens that fail to meet this threshold, deeper levels from the paraffin tissue block may yield adequate cellularity or an alternative specimen may be evaluated if available. If the initial specimen is negative, consideration should be given to repeating the test on another specimen if available.
Testing of the primary tumor versus metastatic sites was also considered by the EP. The available evidence in the literature demonstrates that the TPS scores obtained from the primary tumor and synchronous metastatic lesions are discordant at both the 1% and 50% cutoffs in resected specimens approximately 20% to 30% of the time.59,61,64,65 There were no studies evaluating the RR to PD-1 or PD-L1 ICIs, based on testing specimens from different sites available for review; therefore, the EP could not make a recommendation regarding the testing of a primary tumor versus a synchronous metastatic lesion. When synchronous specimens from multiple sites are available, a reasonable approach would be to test the specimen of the best technical quality (ie, highest viable cellularity and best preservation).
Another important consideration in specimen selection for PD-L1 IHC testing is the potential for discordant results when testing small biopsy samples versus resection specimens. Intratumoral heterogeneity of PD-L1 expression is frequently seen in resection specimens; therefore, it is possible that results obtained from small biopsy samples may differ depending on what region of a tumor is sampled. Retrospective evidence95–97 suggests there is potential variation in the degree of PD-L1 expression in different tumoral patterns, in both lung adenocarcinoma and lung squamous cell carcinoma.95 As a good practice, the EP suggests interpreting PD-L1 TPS with caution when performed on a primary lung biopsy specimen that may contain limited representation of the overall tumor growth, such as a biopsy specimen containing only lepidic component. Given that patients with advanced NSCLC who are candidates for ICI therapy will, in most cases, not undergo resection of the primary tumor or a metastatic lesion, the performance characteristics of PD-L1 IHC using more limited samples (eg, core biopsies, cytology specimens, and endobronchial biopsies) versus resected specimens can only be extrapolated from retrospective analyses of predominantly early-stage tumors. The evidence demonstrates that concordance between small biopsy samples and resected specimens ranges from 83% to 95% at the 50% cutoff and 77% to 86% at the 1% cutoff.82–84 In most discordant cases, a lower score is obtained from the small biopsy sample than from the corresponding resection specimen.98 Based on these findings, testing of small biopsy samples is acceptable.
Many patients with NSCLC, particularly those with advanced disease, are diagnosed by fine-needle aspiration or other cytologic techniques, and such specimens may be the only available pathology material. Studies comparing PD-L1 expression on tumor cells in cytology cell blocks from various sources—including both needle aspirate specimens and body fluids—to expression in paired resection specimens67,70,74,75,77,79,80 show good concordance, ranging from 67% to 94% with a pooled concordance of 85% (95% CI, 0.80–0.90)99 (Figure). A small number of studies have evaluated the use of PD-L1 immunocytochemistry performed on alcohol-fixed cytology material. Noll et al72 and Lozano et al78 compared TPS obtained from alcohol-fixed aspirate slides to that obtained from concurrent histologic core needle biopsies and found an overall concordance rate of 97% (Lozano) and 89% (Noll) at the clinically relevant cutoffs of 1% and 50% (Figure). Of note, Noll et al72 used the 22C3 Dako/Agilent pharmDx clone, while Lozano et al78 used both the VENTANA PD-L1 (SP263) assay (Roche Diagnostics, Indianapolis, Indiana) and the 22C3 clone in their series of cases. Munari et al71 compared PD-L1 scores obtained from ex vivo specimen aspirates and paired whole sections from resections, using the VENTANA BenchMark ULTRASP263 clone, and reported a concordance rate of 81% at the 1% cutoff and 91% at the 50% cutoff (Figure). In most discordant cases, the score obtained from the aspirate smear was lower than that from the paired resection specimen. Of note, both the SP263 and 22C3 assays designate that they are clinically validated for use in formalin-fixed, paraffin-embedded (FFPE) material. While the limited data currently available for review appear promising, at the current time, alcohol-fixed cytologic material should be used for PD-L1 testing only if FFPE materials are not available, and in laboratories that have specifically validated this sample type, and the results should be reported with a comment that PD-L1 testing in non-FFPE tissues has not been clinically validated.
Previously collected archival tissue is often the most readily available source for biomarker testing and mitigates the need for rebiopsy in patients who may not be amenable to further intervention. However, some studies have raised issues regarding antigen decay in archival tissue, particularly nuclear and membranous staining,57,100,101 which may have implications for PD-L1 TPS. In fact, preliminary studies suggest that the PD-L1 clone 22C3 is prone to deglycosylation and diminished expression after a year or more in storage.102 Additionally, tumor expression of PD-L1 is dynamic, unlike other predictive biomarkers, with some discordance in expression reported following treatment between initial and recurrent disease.103 In fact, the study of 2 doses of pembrolizumab (MK-3475) versus docetaxel in previously treated participants with non–small cell lung cancer, or KEYNOTE-010,104 was amended to require PD-L1 evaluation in newly collected samples except when patient safety posed a significant risk. As such, formal evaluation of the utility of archival tissue in this context is of practical importance. At this time, data are limited to a study by Herbst et al,22 which compared PD-L1 status and response rate in archived samples and fresh biopsy samples, using the 22C3 pharmDx assay. In this study, 455 archival samples were evaluated with a median time between sample collection and PD-L1 testing of 250 days (range, 3–2510 days). Five hundred seventy-eight “new” samples were evaluated that had an average of 11 days between sample collection and PD-L1 testing (range, 1–371 days). The results of this study showed that PD-L1 expression levels with TPS of 1% or more and TPS of 50% or more were similar between archival and newly collected samples. The authors22 concluded that PD-L1 expression was adequately preserved following a median 8 months of storage and that expression was potentially not affected by intervening treatment. Additionally, OS and progression-free survival hazard ratios (HRs) were similar across archival and newly collected samples in both 1%-or-more and 50%-or-more TPS cohorts.
Bone is a frequent site of metastasis for lung cancer; hence, the impact of decalcification methods on PD-L1 expression is of key interest. Three studies50,105,106 evaluated the expression of the 22C3 Dako/Agilent pharmDx assay in decalcified tissue samples; Forest et al50 also included the E1L3N (PD-L1 [E1L3N] XP, Cell Signaling Technology, Danvers, Massachusetts) clone. Strickland et al105 evaluated PD-L1 expression following decalcification with EDTA, formic acid/MasterCal IM Plus (FA/MC), 12% hydrochloric acid (HCl), and Decal STAT Decalcifier/23% HCl solutions at periods of 1, 2, 6, and 24 hours. This study found that EDTA and FA/MC had little effect on PD-L1 expression, whereas 12% HCl resulted in a progressive decline in expression. Notably, Decal STAT dramatically reduced expression at all treatment durations. Forest et al50 evaluated PD-L1 expression following EDTA or acid decalcification and did not detect a statistically significant difference in reactivity with the E1L3N clone for the first 24 hours or after delayed fixation following EDTA decalcification. Labeling of carcinoma cells by the 22C3 pharmDx assay was slightly decreased following EDTA decalcification (an observation that did not reach statistical significance); in contrast, there was a significant decrease in percentage of tumor cell staining following acid decalcification. Pontarollo et al106 evaluated PD-L1 expression following EDTA or formic acid decalcification and did not detect a significant difference in TPS in decalcified versus nondecalcified specimens. Available data suggest that specimens decalcified in EDTA or formic acid for less than 24 hours may generate PD-L1 TPS scores that are comparable to those of specimens exposed to formalin alone; however, nondecalcified samples should be prioritized for PD-L1 testing, when available, and laboratories should specifically validate decalcified specimens if they offer testing for this sample type. The use of decalcified specimens for PD-L1 testing to select patients for ICI therapy has not been clinically validated.
The public comment period recorded 78 responders, of whom 66 (84.62%) agreed or agreed with suggested statement modifications, 6 (7.69%) disagreed, and 6 (7.69%) were neutral. There were 10 written comments, many of which suggested that the final recommendation language be changed from “clinician” to “pathologist.” Suggested comments also included a request to clarify the definition of “best available specimen.” All comments were taken into consideration. The recommendation statement was edited to reflect that the pathologist is the intended subject of the recommendation. In addition, the statement was modified to remove language around the best available specimen and instead emphasizes that validation needs to occur for all potential specimen types and fixatives. The discussion text includes an explanation of the variables that make defining an optimal specimen for testing difficult, and the discussion text includes an explanation of the best available specimen, as well as a recommendation to use available guidelines for appropriate validation.
3. When feasible, pathologists should use clinically validated PD-L1 immunohistochemistry assays as intended.
(Strength of Recommendation: Conditional; Certainty of Evidence: Very Low)
The evidence base supporting this statement includes 54 studies reporting on PD-L1 status concordance,37,38,74,78,88,107–120 diagnostic test characteristics,108 and interobserver agreement112,116,121–126 of various combinations of clinically validated PD-L1 assays. Additionally, clinical trials and observational studies reporting on immunotherapy survival and RR,10–38,104,119,127 leading to the clinical validation of the assays, were used as indirect evidence to support this statement. Certainty of evidence for the outcomes was assessed as low and very low. The EP members discussed clinically validated, FDA-approved PD-L1 IHC assays (CDx assays for the purpose of this document) versus laboratory-developed and validated PD-L1 IHC assays at length. From the available evidence, the EP members determined that use of CDx assays carried moderate benefits; however, the harms of CDx assays, including availability of the testing platforms and clones and specific training for each CDx, were also defined as moderate. The EP members concluded that the balance of effects did not favor either CDx assays or laboratory-developed tests (LDTs). The EP members also determined that use of CDx assays carried a large cost and could lead to reduced health equity. A conditional strength of recommendation was based on the clinical validation of these assays and their established ability to predict immunotherapy response but with an understanding of the limitations of this guidance. EP members agreed that the guidance would probably be acceptable to key stakeholders and probably feasible to implement. Refer to Supplemental Tables 5 through 8 for a summary of the risk of bias assessment for all included studies and the certainty of evidence assessment for all outcomes informing the statement, Supplemental Table 3 summarizes the EtD framework.
For PD-L1 testing, there are 4 CDx assays—PD-L1 IHC 22C3 Dako/Agilent pharmDx, PD-L1 IHC 28-8 Dako/Agilent pharmDx, VENTANA PD-L1 SP142, and SP263 assays—each of which was codeveloped with a specific anti–PD-1/PD-L1 agent used in clinical trials and has been approved by the FDA and/or other regulatory agencies as companion diagnostics for the use of those agents (for intended use) in NSCLC. Each CDx assay is determined by a specific combination of an antibody clone, IHC testing platform, and reagents (Table 4). Thus, any PD-L1 IHC assay that consists of different combinations of those of the 4 FDA-approved assays, or (as noted earlier) is used on tissues fixed in a manner not included in the complementary clinical trial(s), is considered an LDT (see Recommendation 4). For financial and/or practical reasons, it may be difficult for a given pathology laboratory to be equipped with multiple CDx assays for PD-L1 testing, and the laboratory may opt to use 1 or a few assays, either CDx or LDTs, to determine the eligibility for treatment with various anti–PD-1/PD-L1 agents.128,129
In this context, it is important to note that the analytic compatibility and exchangeability of PD-L1 IHC assays for tumor cell testing, in particular those of the 4 CDx assays, have been extensively studied.37,38,74,78,88,107–120 Most studies reported similar results—expression concordance between 22C3, 28-8, and SP263 CDx assays is typically 90% or greater, while the VENTANA PD-L1 SP142 CDx generally shows weaker expression than the other 3 CDx assays with fewer positive tumor cells37,115 (Tables 5 and 6). As for reproducibility, interobserver agreements around tumor cell scoring of PD-L1 are similar among the 4 CDx assays and generally substantial to almost perfect for both 1% and 50% cutoffs,112,116,121–126 whereas those on immune cells are only slight or fair.125,126 The technical performance and interobserver agreements of LDTs will be discussed in Recommendation 4.
PD-L1 Status Concordance in FDA-Approved Companion Diagnostic Assays and Laboratory-Developed Tests (LDTs)

Reported Kappas for PD-L1 Status in FDA-Approved Companion Diagnostic Assays and Laboratory-Developed Tests (LDTs)

Fitzgibbons et al57 state that, for initial validation of every assay used clinically, laboratories should achieve at least 90% overall concordance between the new assay and the comparator assay or expected results. Given its failure to meet the guideline recommendations, the SP142 CDx is not considered appropriate for the selection of patients for treatment with pembrolizumab, cemiplimab, or nivolumab/ipilimumab. While data suggest that 22C3, 28-8, and SP263 CDx assays may be interchangeable, information about predictive performance of those CDx assays for anti–PD-1/PD-L1 agents outside of their given indication or that of LDTs is limited.37 The same is true of LDTs, based on the primary antibodies in these CDx assays and on other anti–PD-1/PD-L1 clones. If feasible, pathologists should use CDx assays as intended, until larger cohort studies confirm the comparability of those CDx assays from both technical and clinical standpoints. The FDA has embraced more generic language in certain therapeutic approvals,130 indicating that PD-L1 status “as determined by an FDA-approved test” may be used for patient selection. The laboratory director should be familiar with these testing indications and approvals when determining which assays to validate for clinical use. The CDx assays if used outside of their trial-defined indications require that they undergo a technical validation on par with any LDT as defined by the CAP validation guidelines.
During the public comment period, of a total of 79 responses, 70 (88.61%) agreed or agreed with suggested modifications to the statement, 4 (5.06%) disagreed, and 5 (6.33%) were neutral. There were 28 written comments. Some comments suggested using the term pathologist instead of laboratories for clarity. In addition, there were several comments about the use of the text “clinically validated” in the statement. While the final statement did not reflect this suggested change, this is discussed within many sections addressing validation. The public commented on the importance of clinically validated assays and their utility as intended by the assay manufacturer. Although validation is discussed in the text, the details on how to validate assays and general quality assurance measures are out of scope for this guideline. All comments were taken into consideration. The final recommendation statement was edited to clarify that it is the pathologist carrying out the recommendation.
4. Pathologists who choose to use laboratory-developed tests for PD-L1 expression should validate according to the requirements of their accrediting body.
(Strength of Recommendation: Strong; Certainty of Evidence: Very Low)
The evidence base informing this statement is composed of 1 study reporting on immunotherapy RRs using clones 22C3 and 73-10,131 and 8 studies reporting on PD-L1 status concordance,73,124,131–136 diagnostic test characteristics,124 and interobserver agreement124,125 of LDTs when compared with clinically validated IHC assays. The certainty of evidence was low and very low for the outcomes of interest. From the available evidence, however, EP members defined the benefits of validating all LDTs as large and the harms of the validation as trivial and thus concluded that the benefits outweighed the harms. It is expected that this guidance will be acceptable to key stakeholders and feasible to implement. In addition, given the possible variable performance of unvalidated LDTs when compared to CDx-based PD-L1, the EP members concluded that not performing required validation137 on LDTs could lead to substantial harms to patients, leading to this recommendation as strong, despite a certainty of evidence rating of very low. Refer to Supplemental Tables 5 through 8 for a summary of the risk of bias assessment for all included studies and the certainty of evidence assessment for all outcomes informing the statement. Supplemental Table 3 summarizes the EtD framework.
An LDT is defined as an in vitro diagnostic test that is designed and used within a single laboratory.138,139 The FDA notes that LDTs are important to the continued development of personalized medicine, and that in vitro diagnostics should be accurate so that patients and health care providers do not seek unnecessary treatments, delay needed treatments, or become exposed to inappropriate therapies. An LDT may use one of the available anti–PD-L1 clones used for CDx or another clone (such as E1L3N) that has never been validated in prospective clinical trials, in combination with any available IHC platform128 (Tables 7 and 8). Laboratories may run LDTs for financial or practical reasons, but as the FDA noted, it is important that they be accurate. Thus, all LDTs should be validated. LDTs including those with clone E1L3N can reasonably match the technical performance of CDx assays,73,124,131–133,135,136,140 but multi-institutional studies evaluating a large number of LDTs have reported that only 50% to 60% were analytically compatible to CDx assays.134,141 As for reproducibility, interobserver agreement on tumor cell scoring of PD-L1, using clone E1L3N LDTs among multiple pathologists, was similar to that using 22C3, SP263, and SP142 CDx assays for both 1% and 50% cutoffs.124,125 While specific validation requirements are out of scope with this guideline, technical validation of an LDT using a CDx as the gold standard142 is important for the appropriate selection of patients for treatment with PD-1 axis blockade.
The PD-L1 IHC 73-10 Dako/Agilent pharmDx assay, a combination of the clone 73-10 and a Dako/Agilent platform, has not yet been approved by the FDA, but it was codeveloped with an anti–PD-L1 agent, avelumab, and has been evaluated in multiple clinical trials of patients with NSCLC.143,144 Therefore, the 73-10 IHC assay is distinct from other LDTs, since it is used globally beyond a single laboratory. Data on the analytic performance of the 73-10 IHC assay are scarce, likely owing to its limited accessibility.128 Reportedly, the 73-10 assay is more sensitive than the 4 CDx assays.126 The 73-10 pharmDx assay was optimized to identify tumors with low PD-L1 expression, susceptible to antibody-dependent cellular cytotoxicity through a wild-type Fc region—a unique property of avelumab along with PD-1 axis blockade.143 A study with 231 NSCLC tissue samples, including 83 from a clinical trial for avelumab, showed overall similar analytic and predictive performances along with a higher negative predictive value of the 73-10 assay as compared to those of the 22C3 CDx assay when the cutoffs of 1%, 50%, and 80%, and the cutoffs of 1%, 20%, and 50%, respectively, were compared.131,145 Given the differences in analytic cutoffs used to develop 73-10, compared to those for the other CDx assays, additional studies are warranted to evaluate the analytic compatibility of these assays more carefully.
During the public comment period, of a total of 76 responders, 66 (86.84%) agreed or agreed with suggested modifications to the draft statement, 3 (3.95%) disagreed, and 7 (9.21%) were neutral. There were 10 written comments; some agreed that adequate validation is a must when using an LDT assay, while there were some who advocated for the use of only FDA-approved assays. In addition, there were comments about the low certainty of evidence with a strong recommendation. The final recommendation was modified to reinforce that LDTs are appropriate to use if the assays are adequately validated according to the laboratory’s accreditation requirements. To address the strong recommendation for low levels of certainty of evidence, the EtD framework used by the EP further reinforced the importance of validation with the use of LDT assays.
5. Pathologists should report PD-L1 immunohistochemistry results using a percentage expression score.
(Strength of Recommendation: Conditional; Certainty of Evidence: Very Low)
The guideline statement is supported by 5 studies reporting on immunotherapy RRs17,22,29,31,127 and survival rates17,22,29,31,127 stratified by specific PD-L1 TPS thresholds. An additional 5 studies evaluated interobserver agreement, using multiple IHC clones also stratified by TPS score.112,116,122–124 The certainty of evidence for RRs was assessed as low, based on a very serious aggregate risk of bias across the studies reporting on the outcome, while the certainty for survival and interobserver agreement was assessed as very low owing to very serious risk of bias plus downgrading for inconsistency (Supplemental Table 9). From the available evidence, EP members concluded that reporting PD-L1 expression as a percentage score carried moderate benefits and only small harms, leading to the determination that benefits probably outweighed the potential harms. This recommendation is expected to be acceptable to key stakeholders and will be feasible to implement. This guidance is expected to have no impact on health equity, and the resource requirements were considered to be negligible. Refer to Supplemental Tables 5 through 9 for a summary of the risk of bias assessment for all included studies and the certainty of evidence assessment for all outcomes informing the statement. Supplemental Table 3 summarizes the EtD framework. The levels of PD-L1 TPS demonstrate statistically significant correlation with patient response and survival following immunotherapy with ICI administered in isolation and in chemo-immunotherapy combinations. Published evidence demonstrates higher RR and improved outcomes in patients with higher levels of PD-L1 expression, particularly among those patients with high PD-L1 TPS (at least 50%), both in the context of retrospective correlative studies using “real-life data” and in the context of RCTs.21,46,90,104,146–148 For this reason, the EP recommends that pathologists report PD-L1 IHC results by using a percentage expression score.
From a practical standpoint, providing an exact percentage expression score value may be challenging owing to the subjective nature of visual assessment of PD-L1 expression and scoring variability among pathologists, particularly regarding assessment of immune cell populations. Recent publications have shown high interpathologist concordance when scoring tumor cells, but lower agreement on immune cell scoring across different PD-L1 assays.90,125,127 In addition, other PD-L1 test–related variables such as sample selection and availability, differences between PD-L1 assays, and intratumor heterogeneity of PD-L1 expression may influence the ability of pathologists to consistently provide precise quantitative measurements. For this reason, one option that several clinical laboratories have adopted is to report ranges of PD-L1 percentage expression scores (eg, 5% or 10% incremental values) instead of absolute scores. This semiquantitative approach is expected to be more accurate and reproducible than reporting specific expression percentage values and still provides sufficient information for management decisions, provided that ranges are reported with management-based cut points in mind. A more objective quantification of PD-L1 expression could conceivably be achieved with the use of assay-specific controls representing expression intensities across the dynamic range of the test.105
A total of 77 responses were received during the public comment period, of which 65 (84.42%) agreed or agreed with suggested modifications to the draft statement, 3 (3.90%) disagreed, and 9 (11.69%) were neutral. There were 27 written comments. The most common comments to the draft statement included using percentage cutoffs by assay, and how scoring is dependent on a specific clone. The final recommendation was slightly edited without change to the context. All comments received were reviewed and addressed with the discussion.
6. Clinicians should not use tumor mutation burden alone to select patients with advanced NSCLC for immune checkpoint inhibitors, based on insufficient evidence in this population.
(Strength of Recommendation: Conditional; Certainty of Evidence: Very Low)
The evidence base is composed of 8 studies reporting on immunotherapy RRs149,150 and survival rates29,32,35,149–153 when correlated with TMB status. Certainty of evidence was assessed as very low for RRs and very low for survival rates. This conditional recommendation was based on the trivial benefits of using TMB to select patients for immunotherapy paired with the moderate harms of its use. The balance of effects favored not using TMB, as did the moderate costs and probable reduced health equity that would be associated with recommending its use. Further to these domains, the EP members concluded that guidance in support of TMB would not be acceptable to key stakeholders and probably not be feasible to implement. Refer to Supplemental Tables 5 through 8 for a summary of the risk of bias assessment for all included studies and the certainty of evidence assessment for all outcomes informing the statement. Supplemental Table 3 summarizes the EtD framework.
Recommendation statement 6 regarding the use of TMB received a conditional strength based on low and very low certainty of evidence for impact on survival and RRs, respectively. Regarding the line of therapy that is relevant to this recommendation, the 8 studies evaluated in this evidence base include heterogeneous lines of administration of therapy, including in the context of first-line therapy152,153 ; a combination of lines from first through third35,149 ; palliative setting29 ; and undisclosed treatment setting.32,151
Despite initial findings from the CheckMate-22716 trial showing an improved 1-year progression-free survival rate for patients with advanced NSCLC and TMB of 10 mutations/megabase (mut/Mb) or greater and receiving nivolumab plus ipilimumab in the first-line setting,154 the supplemental FDA application was withdrawn when subsequent data showed no difference in survival outcomes between patients stratified by high or low tumor TMB. The median OS with nivolumab + ipilimumab in patients with TMB of 10 mut/Mb or greater was 23.03 months versus 16.72 months for the chemotherapy arm (HR, 0.77; 95% CI, 0.56–1.06); among patients with TMB lower than 10 mut/Mb, the median OS was 16.20 months versus 12.42 months, respectively (HR, 0.78; 95% CI, 0.61–1.00). At present, the FDA approval for nivolumab plus ipilimumab as first-line treatment for patients with metastatic NSCLC requires tumor PD-L1 expression of 1% or more, as determined by an FDA-approved test, with no EGFR and ALK genomic tumor aberrations, and does not include TMB.
Accelerated approval was granted for pembrolizumab in the treatment of adult and pediatric patients with unresectable or metastatic TMB-high (≥10 mut/Mb) solid tumors—as determined by an FDA-approved test—who have progressed following prior treatment and who have no satisfactory alternative treatment options. Although this is a tumor-agnostic approval, it is worth noting that the KEYNOTE-158 study155 (the basis of this approval) included 102 patients with TMB of 10 mut/Mb or greater spanning 9 different tumor types, none of which were NSCLC.156 These data came from an analysis of 10 cohorts of patients with various previously treated, unresectable or metastatic solid tumors, and response rates (not survival rates) were compared against those who did not have high TMB.
A retrospective study157 evaluating 909 nonsquamous NSCLCs, with both PD-L1 and TMB data available, identified that the median TMB was significantly higher in the PD-L1–high group than in the PD-L1–low and negative groups (median, 12.2 mut/Mb versus 10.6 mut/Mb versus 10.6 mut/Mb, respectively; P < .001) with a modest, but significant linear correlation between TMB and PD-L1 TPS (P = .12; P < .001). Additionally, in a multivariable logistic regression analysis evaluating factors associated with response to ICIs, improved progression-free survival was significantly associated with both higher PD-L1 expression (HR, 0.41 [95% CI, 0.28–0.62]; P < .001) and TMB (HR, 0.97 [95% CI, 0.95–0.99]; P = .002), whereas OS was significantly associated only with PD-L1 expression (HR, 0.59 [95% CI, 0.38–0.93]; P = .02).
The open comment period collected a total of 78 responders, of whom 59 (75.64%) agreed or agreed with suggested modifications to the draft statement, 4 (5.13%) disagreed, and 15 (19.23%) were neutral. There were 12 written comments. These included the recent FDA approval of pembrolizumab with the use of TMB-high. Other comments included the use of TMB-high only if treatment options are not available. The EP reviewed all comments and decided that no edits were necessary, although the FDA approval should be part of the discussion text.
LIMITATIONS
PD-L1 expression and TMB testing are the most widely used biomarkers in patients with NSCLC being considered for ICI therapy. Despite widespread adoption, findings from the extensive literature review informing this guideline demonstrate that conclusions about several PD-L1 and TMB test-related variables are difficult to draw owing to the limited number of publications addressing specific technical aspects associated with these biomarkers. Notwithstanding these limitations and the complexity surrounding testing of these biomarkers, the EP developed evidence-based recommendations that address preanalytic, analytic, and postanalytic considerations for PD-L1 and TMB testing in the clinical setting.
FUTURE DIRECTIONS
PD-L1 is widely recognized as an imperfect biomarker for the selection of patients for ICI therapy. While many patients with NSCLC are now receiving ICI therapy at some stage of their care and recent improvements in cancer outcomes are unparalleled in modern oncologic history, most patients with NSCLC will not derive a response or survival benefit. Beyond lack of response, some patients have severe adverse effects related to aberrant immune activation.158,159 Analyses of the tumor microenvironment (TME) and host factors have identified other considerations that may influence response to ICI therapy.1 Relatively simple variables such as CD8+ T-cell enumeration in the TME may add to the predictive power of PD-L1 status, but this approach has not been widely validated in clinical practice.95,160 More complex analyses, including multiplex immune profiling by IHC or immunofluorescence to capture both the PD-L1/PD-1 status and immune cell characterization in the TME, have been shown to outperform PD-L1 IHC alone. While these strategies have been piloted in a number of academic centers,161,162 complexities of image analysis and data storage have slowed adoption beyond large academic centers. Other conceptually straightforward analyses such as PD-1/PD-L1 proximity may be superior to PD-L1 status for predicting ICI response, but are not widely available.163 Host factors such as gut microbiome, human leukocyte antigen (HLA) genotype, and neutrophil to leukocyte ratio have all been examined extensively as factors influencing ICI therapy response and are increasingly incorporated into clinical trials either as components of the primary design or as correlative biomarkers.164 As highlighted by the widespread interest in TMB as an ICI therapy biomarker, other specific tumor genome characteristics beyond driver oncogene mutation status are subject to extensive study in ICI-treated patient cohorts. Comprehensive characterization of the tumor genome, using next-generation sequencing technology, is required to capture the suite of co-mutations (such as in serine/threonine kinase 11 [STK11], Kelch-like ECH-associated protein 1 [KEAP1], switch/sucrose non-fermentable SWI/SNF pathway genes), mutational signatures (tobacco, apolipoprotein B mRNA editing enzyme, catalytic polypeptide [APOBEC], microsatellite instability, homologous recombination deficiency), and genome state changes165 that may inform ICI treatment outcomes.166 Given the complexity of the tumor-immune interaction, biomarker development is likely to benefit from artificial intelligence algorithms designed to improve the quantification and synthesis of pathologic, genomic, radiomic, and clinical data.162,167,168 Access to this level of information is currently limited to a small subset of patients with cancer globally. While these types of advanced analyses are essential to drive discovery, it remains of paramount importance to identify easily implemented and cost-effective biomarkers that can identify patients most likely to benefit (or not benefit) from ICI therapy.
CONCLUSIONS
This guideline was developed during the course of 4 years, during which time the use of immunotherapy for patients with lung cancer gained greater traction, and our understanding of the factors that contribute to RRs grew in sophistication. That said, PD-L1 IHC testing remains a cornerstone of NSCLC biomarker testing, and despite its less-than-ideal negative and positive predictive values, most patients with advanced NSCLC will have their tumors tested for PD-L1 expression.169 The development of different paired diagnostic assays for each of the immunotherapy drugs taken through clinical trials to regulatory approval has created a confusing landscape of companion (on-label “required”) diagnostics and complementary (on-label “nice to know”) diagnostics, along with a plethora of PD-L1 antibody clones and testing platforms. The EP recognized that the regulatory-approved diagnostics are clinically validated, and as such their use is recommended. However, the panel also recognized that the practical reality of most laboratories—including lack of access to the full suite of approved clones and platforms and the increased cost of running CDx-labeled assays—may require use of LDTs. To ensure patient access to PD-L1 testing, particularly at the local level, this panel also endorses the use of LDTs following technical validation against 1 or more of the approved CDx PD-L1 assays. Formal IHC validation recommendations are beyond the scope of this guideline; the reader is directed to IHC validation guidelines published by the CAP. It is incumbent on testing laboratories to recognize the biological and technical variables that influence PD-L1 expression status, including its expression heterogeneity, the quantitative sample requirements, and appropriate validation of the chosen assay for each of the sample types that may be used to render a diagnosis in patients with lung cancer. It is also important to recognize that other factors may contribute to the decision to proceed with ICI therapy in patients with NSCLC, including the presence of genomic driver alterations such as in EGFR and ALK, suggesting a lower efficacy of ICI. TMB has been proposed as a pan-cancer biomarker of ICI response, but there is a dearth of published data to date suggesting that the current cut point defining TMB-high is a reliable predictor of response to ICI in patients with NSCLC.
Guideline Revision
This guideline will be assessed every 5 years to determine if an update is warranted. When appropriate, the panel may recommend an earlier update to the CAP in collaboration with ASCO, AMP, IASLC, PPS, and the LUNGevity Foundation.
Disclaimer
The CAP developed the Pathology and Laboratory Quality Center for Evidence-Based Guidelines as a forum to create and maintain laboratory practice guidelines (LPGs). Guidelines are intended to assist physicians and patients in clinical decision-making and to identify questions and settings for further research. With the rapid flow of scientific information, new evidence may emerge between the time an LPG is developed and when it is published or read. LPGs are not continually updated and may not reflect the most recent evidence. LPGs address only the topics specifically identified therein and are not applicable to other interventions, diseases, or stages of diseases. Furthermore, guidelines cannot account for individual variation among patients and cannot be considered inclusive of all proper methods of care or exclusive of other treatments. It is the responsibility of the treating physician or other health care provider, relying on independent experience and knowledge, to determine the best course of treatment for the patient. Accordingly, adherence to any LPG is voluntary, with the ultimate determination regarding its application to be made by the physician in light of each patient’s individual circumstances and preferences. CAP makes no warranty, expressed or implied, regarding LPGs and specifically excludes any warranties of merchantability and fitness for a particular use or purpose. CAP assumes no responsibility for any injury or damage to persons or property arising out of or related to any use of this statement or for any errors or omissions.
The authors thank the collaborating societies and their staff involved in the development of this guideline: American Society of Clinical Oncology, Association for Molecular Pathology, International Association for the Study of Lung Cancer, Pulmonary Pathology Society, and LUNGevity Foundation. The authors also gratefully acknowledge advisory panel members for their careful review and guidance throughout the development of the guideline and for their thoughtful review of this work: Ezra Baraban, MD, Eric Bernicker, MD, Russell Broaddus, MD, PhD, Sanja Dacic, MD, PhD, Fang Fan, MD, PhD, Patrick Fitzgibbons, MD, Zaibo Li, MD, PhD, Robert McGee, MD, PhD, Sinchita Roy-Chowdhuri, MD, PhD, Marina Vivero, MD, Barbara A. Ward, BA, Ahmet Zehir, PhD; and Carol F. Colasacco, AHIP, MLIS, SCT(ASCP), Nicole Thomas, MPH, CT(ASCP), and Marisol Hernandez, MLS, MA for their support throughout the guideline development process.
References
APPENDIX
Disclosed Interests and Activities From July 2018 to October 2023
Mark Awad, MD, PhD
Consulting fees or Advisory Board: Affini-T Therapeutics Inc, AstraZeneca, Bristol-Myers Squibb, EMD Serono Inc, Genentech, Gritstone bio, Foundation Medicine, Instil Bio, Janssen Oncology, Merck, Mirati Therapeutics Inc, Novartis, Pfizer, Regeneron Pharmaceuticals
Research grants: AstraZeneca, Bristol-Myers Squibb
David M. Hwang, MD, PhD
Consulting fees or Advisory Board: Amgen Inc, Bayer Canada, GSK Canada, Merck Canada Inc, Novartis Canada, Roche Canada
Research grants: Boehringer Ingelheim
Speakers’ bureau/lecture fees/honoraria: Amgen Canada Inc, AstraZeneca Canada, Merck Canada Inc, Pfizer Canada Inc
Gregory Kalemkerian, MD
Research grants: Blueprint Medicines, Cullinan Oncology, Daiichi Sankyo, Merck, Takeda Pharmaceuticals
Fernando Lopez-Rios, MD, PhD
Consulting fees or Advisory Board: AstraZeneca, Bayer, Bristol-Myers Squibb, Daiichi Sankyo, Janssen Oncology, Eli Lilly, Merck Sharp & Dohme Corp, Roche
Research grants: Eli Lilly, Pfizer, Roche Thermo Fisher Scientific
Speakers’ bureau/lecture fees/honoraria: AstraZeneca, Bayer, Bristol-Myers Squibb, Eli Lilly, Janssen Oncology, Merck Sharp & Dohme Corp, Pfizer, Roche, Takeda Pharmaceuticals, Thermo Fisher Scientific
Mari Mino-Kenudson, MD
Consulting fees or Advisory Board: AstraZeneca, Janssen Oncology, Sanofi
Lauren Ritterhouse, MD, PhD
Employment: Foundation Medicine*
Speakers’ bureau/lecture fees/honoraria: Astellas Pharma Inc, EMD Serono Inc, Sanofi, Thermo Fisher Scientific
Stock options/bonds: Foundation Medicine*
*Employment and stock options were obtained after completion of the guideline manuscript.
Lynette M. Sholl, MD
Consulting fees or Advisory Board: AstraZeneca, Eli Lilly, Genentech
Research grants: Bristol-Myers Squibb, Genentech
Paul E. Swanson, MD
Consulting fees or Advisory Board: Wolters-Kluwer Publishing
No Disclosures to Report
Upal Basu Roy, PhD, MS, MPH, Mary Beth Beasley, MD, Richard Walter Cartun, PhD, MS, Larissa V. Furtado, MD, Ajit Paintal, MD, Kearin Reid, MLS(ASCP), MLIS, Lesley A. Souter, PhD, Christina B. Ventura, MPH, MT(ASCP)
Author notes
Supplemental digital content is available for this article at https://meridian.allenpress.com/aplm in the July 2024 table of contents.
Cartun is now retired. Sholl and Furtado are the Guideline Expert Panel Co-chairs
Competing Interests
Authors’ disclosures of potential conflicts of interest and author contributions are found in the Appendix at the end of this article.