Neoadjuvant Therapy and Lung Cancer: Role of Pathologists

The gross examination and tumor sampling for microscopic evaluation is the first important step in the evaluation of surgically resected lung cancers after neoadjuvant therapy. Adequate sampling is essential for the calculation of the residual viable tumor, which includes relation to necrosis and stroma within the tumor bed. The accuracy of this assessment depends on sufficiently sampled tumor bed. The most accurate approach would be to submit the entire tumor bed; however, this is impractical, resulting in a large number of slides and time-consuming evaluation for larger tumors. The IASLC multidisciplinary recommendation addressed this issue and proposed a standardized grossing protocol.⁹  The IASLC recommendation is that a tumor bed of 3 cm or smaller should be entirely submitted for microscopic examination. If the tumor bed is greater than 3 cm, the largest cross section should be entirely submitted with additional sections per 1 cm. Histologic sections at the periphery of the tumor bed should include the border of the tumor with at least 1 cm of the surrounding nonneoplastic lung parenchyma to define the edge of the tumor bed. If the initial microscopic evaluation suggests a pCR, the remaining entire tumor bed should be submitted for microscopic review. In addition to the IASLC recommendation, other approaches to grossing have been published recently. One of the proposals for tumors larger than 3 cm was to submit at least 50% of the tumor bed.¹⁰  Weissferdt et al¹¹  showed that the submission of the entire primary tumor up to a maximum of 20 sections is required for the most accurate assessment of a residual viable tumor and pathologic response.

Before grossing of the specimen, it is essential to review pretreatment and the most recent computed tomography scans of the chest, which can help to identify where the tumor bed should be. This is particularly important in cases with pCR or MPR close to 10%, which frequently show a small delicate scar that may be difficult to identify on gross inspection (Figure 1, A through C). In cases with little or no response, the gross appearance of the tumor bed is similar to untreated tumors, such as easily identifiable pleural retraction or tumor mass on cross section (Figure 2). The gross descriptions should contain an estimate of the percentage of necrosis. This is particularly important for the large necrotic cavitary tumors for which the microscopic assessment of necrosis may be inaccurate and correlation between gross and microscopic findings is essential. Dense fibrosis or organizing pneumonia can appear white or tan on gross examination and is difficult to distinguish from viable tumor. In those cases the size of the tumor bed may be overestimated and it is essential to correlate microscopic findings with gross description.

Lymph nodes should be submitted entirely. If the lymph node metastasis measures more than 2 cm, the lymph node can be bisected and the central section through the tumor can be submitted. However, individual laboratories may follow their internal guidelines for lymph node sampling, including sampling of the entire lymph nodes regardless of their size.⁹ 

Microscopic evaluation of the tumor bed includes the quantitative assessment of the viable tumor, necrosis, and stroma in 10% increments for a sum of 100% (Figure 3). If the percentage is less than 5%, single digits should be used. Stroma is histologically composed of fibrosis and areas of inflammation. The same histologic features can be seen in treatment-naïve tumors, and therefore, it is essential for pathologists to be aware of the treatment history. There are no unique morphologic features that distinguish treatment effect from treatment-naïve tumor morphology (ie, stroma).¹²  Parenchymal scars, particularly apical caps, identified in treatment-naïve lung specimens, are indistinguishable from treatment-related fibrosis. Junker et al^13,14  suggested that the histologic changes of therapy response occur at the tumor periphery, whereas in treatment-naïve cases changes usually occur in the center of the tumor. This observation is somewhat controversial, as others observed cases with treatment effect throughout the tumor bed.^15,16  The type of fibrotic response has been reported to be of prognostic significance in some tumors such as melanoma, but in NSCLC only the percentage of residual viable tumor has prognostic significance.^17,18 

Treatment-induced necrosis can be difficult to distinguish from pretreatment tumor necrosis in some cases. A correlation with gross findings is essential to adequately quantify the amount of necrosis. For example, in cases with a cavitary tumor bed, viable tumor areas are usually submitted for microscopic evaluation. In those cases, the percentage of viable tumor can be overestimated if microscopic findings are not correlated with gross findings.

Although the microscopic assessment of the pathologic response is usually straightforward, there are certain features that may cause diagnostic challenges such as the assessment of the tumor bed and distinction between reactive and neoplastic changes. The nonneoplastic lung adjacent to the tumor bed may show reactive changes, such as prominent type II pneumocyte atypia, which is difficult to differentiate from lepidic adenocarcinoma, and no immunohistochemical marker is helpful. In those cases, the best approach is to identify nonneoplastic intact lung alveolar septa, which are not a component of the tumor bed. This should help to compare the degree of cytologic atypia and determine the tumor bed border. Similarly, cicatricial organizing pneumonia can make it challenging to measure the tumor bed. In such cases, correlation with pretreatment computed tomography studies may be helpful. In some cases distinction between histiocytes and tumor cells may be difficult particularly in cases with significant necrosis. Although the IASLC recommendation does not encourage use of immunohistochemistry, in the latter scenario the carcinoma and histiocytic markers could be diagnostically helpful. There are several histologic subtypes of adenocarcinomas for which the assessment of pathologic response has not been standardized by using an evidence-based approach. For example, in mucinous adenocarcinomas, extracellular mucin may be considered to represent either viable tumor or stroma. Mucinous adenocarcinomas usually do not show response to neoadjuvant therapies, and therefore extracellular mucin can be included in the calculation of the percentage of viable tumor. For papillary and lepidic adenocarcinoma, a similar approach can be used. Fibrovascular cores or fibrous septa should be interpreted as viable tumor rather than stroma. The IASLC reproducibility study demonstrated that this approach simplifies scoring and increases the degree of agreement among readers.¹⁶ 

Previously reported studies used different approaches for estimating the percentage of each tumor bed component microscopically. A semiquantitative or “eyeball” approach was used in the early clinical trials that established the prognostic significance of 10% cutoff for MPR.¹³  Other studies used a total percentage of viable tumor by averaging the results across all slides of the tumor bed.^19,20  Recently, an MPR calculator tool was developed that provides a weighted MPR based on the proportions of total tumor bed area on each slide for a case.¹⁰  There is no consensus on which method should be used. The IASLC reproducibility study showed no significant difference in the assessment of MPR between slide average (unweighted method) and MPR calculator (weighted method).¹⁶  The 10% cutoff for major pathologic response for all histologic subtypes of NSCLC has been challenged by several studies.^17,21,22  Qu et al¹⁷  proposed 65% for adenocarcinoma and 10% for squamous cell carcinoma. Similar cutoffs were reported in studies with European and Asian patients, but the recent study by Pataer et al²³  suggests that the 10% cutoff is applicable to all histologies.^21,22 

Two scoring systems have been proposed for the pathologic response assessment in lung cancer specimens after neoadjuvant therapy—IASLC scoring and immune-related pathologic response criteria (irPRC) scoring.^9,24,25  The IASLC scoring system is applicable to lung cancer only and is therapy agnostic. In contrast, irPRC has been proposed for all tumor types treated with immune checkpoint inhibitors. The irPRC scoring system defines the tumor bed as the sum of the residual viable tumor, necrosis, and regression bed. The regression bed is defined as proliferative fibrosis with neovascularization and evidence of immune activation and cell death. The intratumoral stroma is counted as viable tumor. Even though there are some differences between the 2 scoring systems, namely definitions of regression bed and stroma, there are also similarities particularly in the assessment of viable tumor.

There are no specific guidelines for the assessment of treatment response in synchronous multiple tumor nodules, which may represent either separate primary tumors or intrapulmonary metastases. A practical approach for separate primary tumors would be to determine pathologic response in each tumor separately. For intrapulmonary metastasis, 1 combined pathologic response score may be provided. However, it is worth mentioning that this is an area requiring further investigation.

Artificial intelligence (AI)–powered digital pathology may improve pathologic assessment in surgically resected lung cancer specimens after neoadjuvant therapy.^26,27  A convolutional neural network–based model was developed to digitally assess the percentage of viable tumor in whole slide images and determine MPR in the phase II LCMC 3 study, which evaluated preoperative atezolizumab in patients with resectable NSCLC.²⁸  The AI-powered tool showed a strong correlation with pathologists’ MPR assessment under the microscope. However, the AI tool showed better association with longer disease-free survival and overall survival.²⁸ 

The prognostic significance of lymph node status in the neoadjuvant setting has been established in chemotherapy and immunotherapy cohorts.^20–22,29  Pathologic assessment of response in lymph nodes has not been adequately addressed since early chemotherapy trials. Junker et al^13,14  proposed that lymph nodes should be assessed in the same way as the primary tumor bed. It is recommended to report percentage of viable tumor, necrosis, and stroma. The 10% cutoff for viable tumor in the lymph nodes has been recently challenged. Pataer et al²⁹  reported 70% as the optimal cutoff for lymph node pathologic response, while a study in the Asian population reported 8%.²²  Metastatic mucinous adenocarcinoma in the lymph nodes can show only mucin without viable tumor epithelial cells. For those cases, the IASLC recommendation is to interpret a lymph node as negative for metastatic disease.⁹  pCR in a lymph node can be recognized if there is a well-defined scar or tumor necrosis in the absence of viable tumor cells. It is difficult to determine tumor stromal inflammation owing to the background of lymphocytes. Anthracosilicotic nodules and granulomas should not be mistaken for response to therapy. The presence of carbon pigment and polarizable needle-shaped foreign particles, together with a knowledge of pretreatment-positive lymph node stations, can help to make this distinction. Significance of the size of metastasis and microscopic anatomic location (subcapsular versus replacement) has not been extensively reported in the literature and currently it is uncertain whether these findings need to be reported. The total number of positive lymph nodes should be reported in the same way as in the treatment-naïve cases, but staged as ypN.

Neoadjuvant treatment is becoming the standard of care for patients with advanced surgically resectable NSCLC, and pathologic evaluation of treatment response is emerging as an important factor in the immediate assessment of clinical efficacy. Global standardization of grossing and microscopic assessment in clinical trials and routine clinical practice is needed. Digital pathology and AI tools can help standardization, but cost, access to those tools, and their implementation in routine clinical practice represent common challenges. The ongoing clinical trials should provide evidence for recommendations for the assessment of response in lymph nodes and further refinement of the initial practice recommendations for pathologic evaluation of the primary tumor bed. Blood-based assays for the evaluation of residual disease are becoming increasingly important and it remains to be seen how the results of these assays together with assessment of pathologic response will guide patient treatment.