Assessment of Breast Pathology Reporting Needs and Development of Tumor Synoptic Templates in Sub-Saharan Africa Open Access

A needs assessment using mixed qualitative and quantitative methods was conducted at 3 large referral hospitals in SSA: Kilimanjaro Christian Medical Centre in Moshi, Tanzania; Obafemi Awolowo University Teaching Hospitals Complex in Ile Ife, Nigeria; and Maputo Central Hospital in Maputo, Mozambique. First, 30-minute semi-structured interviews were conducted between June 2021 and June 2022 with key stakeholders involved in breast cancer care. All interviews were performed by a breast pathologist with subspecialty training over a teleconferencing platform (Zoom, San Jose, California). The interview was part of a more global needs assessment about breast pathology services and was structured around the question “In your opinion, what would be an ideal pathology service in breast pathology considering the current available resources?” Additional follow-up questions about breast pathology reporting were asked based on the individual’s initial response. Stakeholders included surgeons, radiologists, pathologists, and oncologists.

Responses were recorded, key points were transcribed and coded using thematic content analysis, and major themes were identified. Then, a synoptic template was compiled from publicly available breast pathology reporting templates from the CAP, RCP, and ICCR.^7–9  The participants were then asked to review the synoptic template and classify each data element: essential, optional, or exclude. Each data element was then labeled as essential, optional, or exclude, based on the majority vote. These responses were then used to develop resource-adapted tumor synoptic templates from consensus review by local pathologists, external pathologists, and at least 1 treating physician from each site.

In addition, a retrospective review of breast pathology reports between 2020 and 2022 was conducted at each institution. A randomly selected sample of reports included 10 biopsies and 20 excisions for malignancy (ie, lumpectomies and mastectomies) with a minimum of 10 cases with reported estrogen receptor (ER), progesterone receptor, and human epidermal growth factor receptor 2 (HER2) immunohistochemistry results. Breast pathology reports were compared to the data elements in the resource-adapted tumor synoptic template, and each component was categorized as present or absent. The reports from each institution were reviewed locally and reviewed again independently by a group of external pathologists. Any discrepancies were resolved by using consensus review and discussion between the local pathologists and external pathologists. The mastectomy reports were also evaluated, based on the Breast Cancer Initiative (BCI) 2.5’s stratified recommendations for breast cancer pathology services and reporting in basic, limited, and enhanced resource settings.¹⁰  This study received a waiver from the institutional review boards of each institution participating in >the study (Kilimanjaro Christian Medical Centre; Obafemi Awolowo University Teaching Hospitals Complex; Maputo Central Hospital; and Memorial Sloan Kettering Cancer Center, New York, New York).

Semi-structured interviews were conducted with 17 participants across all 3 hospitals in Tanzania, Nigeria, and Mozambique. Participants consisted of physicians involved in breast cancer care, including surgeons, radiologists, pathologists, and oncologists (Table 1). In this publication, we will focus specifically on identified problems related to pathology reports.

All participants expressed the desire for higher-quality pathology reports and would welcome the standardization of reports. One participant stated: “It would be great to have electronic diagnostic templates to help expedite reporting and to serve as reminders of what to report.” Overall, participants felt that certain data elements should be included in all reports, including pathologic diagnosis, tumor type and grade, size based on both gross and microscopic evaluations, number of lymph nodes, margin status, and upfront biomarker testing. Some surgeons expressed issues of trust when it comes to pathology: “Sometimes surgeons do not trust the pathology report since we clearly see discrepancies between the specimen that was submitted and the pathology report. For example, we clearly see that lymph nodes were submitted with the main specimen and when we receive the report the lymph node count seems to be off.” Therefore, issues were identified in how both the gross and microscopic evaluations were reported. The pathology laboratories also were interested in using electronic synoptic templates as a potential solution to standardize reports.

A synoptic template with 20 data elements for malignant biopsy specimens (Table 2) and 32 data elements for mastectomy specimens (Table 3) was compiled by using publicly available templates from the CAP, ICCR, and RCP. All 17 stakeholders who were interviewed also reviewed the synoptic template and categorized each data element as essential, optional, or exclude. For malignant biopsy specimens, 5 data elements were considered optional (tumor site/location, presence of ductal carcinoma in situ [DCIS], description of DCIS, presence of lymphovascular invasion [LVI], and gross description for core biopsies) and 3 data elements were recommended to be excluded by stakeholders (tumor size on core biopsy, presence of microcalcifications, and microscopic description) (Table 2). For malignant mastectomy specimens, 7 data elements were considered optional (tumor site/location, presence of DCIS, description of DCIS, presence of LVI, size of largest metastatic focus in lymph node, presence of extracapsular extension in lymph node, and treatment effect in lymph node) and 2 data elements were recommended to be excluded (presence of microcalcifications and microscopic description) (Table 3). Locally adapted breast cancer synoptic templates for breast biopsies of malignant masses with 18 essential elements (Table 3) and mastectomies for malignancy with 28 essential and 1 optional data element were developed from consensus discussion (Table 4).

Reports of biopsies of malignant masses—including core needle biopsies and incisional biopsies—were reviewed from site A (n = 14), site B (n = 14), and site C (n = 3). Site C had challenges in identifying biopsies of malignant masses, resulting in a small sample size. Data elements that were found in all reports at all sites were adequate patient identifiers, the date a specimen was received, the date that the final pathology report was finalized, a gross description, and the histologic tumor type (Table 5). Site A included a summary of block designations most often (12 of 14, 86%). Site B was the most consistent in including laterality and procedure type (14 of 14, 100%). Data elements that were not well documented at all sites included tumor site/location within the breast (ie, quadrant or o’clock position), ranging from 0% (0 of 3) of reports at site C to 14% (2 of 14) of reports at site A; presence and type of DCIS, ranging from 0% (0 of 14) at site A to 67% (2 of 3) at site C; presence of LVI, ranging from 0% (0 of 14) at sites A and B to 67% at site C (2 of 3); tumor size, ranging from 0% (0 of 14) at sites A and B to 67% (2 of 3) at site C; and presence of microcalcifications, ranging from 0% (0 of 14) at sites A and B to 33% (1 of 3) at site C. None of the sites used semi-quantitative scoring to report biomarker results (ie, no reporting of the percentage of positive tumor cells and the intensity of staining).

Excisions for malignancy, including lumpectomies and mastectomies, were reviewed from site A (n = 22), site B (n = 20), and site C (n = 20) (Table 5). Of the pathology components identified as critical for breast cancer care during stakeholder interviews (ie, pathologic diagnosis, tumor type, tumor grade, tumor size, number of lymph nodes, margin status, and biomarker testing results or a recommendation to perform biomarker testing), 5% (1 of 22) of mastectomy reports from site A, 30% (6 of 20) from site B, and 55% (11 of 20) from site C contained all elements. None of the pathology reports had all data elements categorized as essential in the modified tumor synoptic template (Table 5). The data elements that were most consistently documented across all sites were adequate patient identifiers, date a specimen was received, date the final pathology report was finalized, a gross description, the clinical history, the organ, procedure type, and histologic tumor type (Table 5). Essential data elements completely absent from all the pathology reports across all sites were the size of the largest metastatic focus in a lymph node, presence of extracapsular extension, and treatment effect on lymph nodes. The size of invasive tumor was inconsistently documented in site A (45%, 10 of 22) and site B (75%, 15 of 20) but was documented in all cases reviewed for site C (100%, 20 of 20). Overall margin status, that is, negative or positive, was reported for most cases for site A (86%, 19 of 22) and site B (85%, 17 of 20) and in all cases for site C (100%, 20 of 20). Specific distance to margins was reported in 5% of cases for site A (1 of 22), 5% of cases for site B (1 of 20), and 95% of cases for site C (19 of 20). Of note, it is not clear how margins were determined because of all reports reviewed, only 1 case in site A reported the use of ink in the gross description (2%, 1 of 62). Major areas of deficiency were seen in data elements related to neoadjuvant treatment, such as the presence of treatment effect, size of the tumor bed, and the presence of treatment effect in the lymph node (Table 5). When evaluating the reports using BCI 2.5’s basic pathology recommendations, 9% (2 of 22) of reports in site A, 50% (10 of 20) of reports in site B, and 65% (13 of 20) of reports in site C had all 5 essential parameters, which consist of tumor size, lymph node status, histologic type, histologic grade, and TNM stage (Table 6). When assessing for pathology elements recommended for a BCI 2.5 limited pathology setting, none (0 of 22) of the reports from site A, 10% (2 of 20) from site B, and 25% (5 of 20) from site C had all 4 suggested data parameters, which include ER status, margin status, presence of LVI, and presence of DCIS. For BCI 2.5 enhanced pathology elements, 36% (8 of 22) of reports from site A, 75% (15 of 20) of reports from site B, and 65% (13 of 20) of reports from site C included both HER2 and progesterone receptor results.

High-quality pathology services are critical for accurate diagnosis as well as for determining prognosis and optimal treatment. However, diagnostic capacity in SSA is insufficient to meet current cancer care needs. We conducted a needs assessment by using qualitative and quantitative methods to identify critical gaps in breast pathology services as well as to create resource-adapted breast cancer tumor synoptic templates. Through semi-structured interviews, we identified that one of the major barriers to high-quality breast pathology services included incomplete and nonstandardized pathology reports.

Based on detailed stakeholder input, locally adapted breast cancer synoptic templates were created and were then used to conduct a concise review of pathology reports across 3 hospitals in SSA. The modifications that were made to the synoptic template reflect the limited resources available and differences in diagnostic practices at the 3 large referral hospitals as compared to health care facilities in high-resource settings. For example, reporting of microcalcifications in biopsies was felt to not be important by stakeholders because of the lack of a national screening program in all 3 countries. In addition, most biopsies are performed in symptomatic patients with palpable breast masses. In Nigeria, Tanzania, and Mozambique, approximately 67% to 91% of women present with stage 3 or 4 disease, which is a trend that is observed throughout SSA.^11–15  A systematic review found that 77% of patients in SSA present with late-stage breast cancer at the time of diagnosis.¹⁶  Moreover, there is a paucity of radiation therapy available that would allow for de-escalated treatment, such as partial mastectomy and sentinel lymph node biopsy. As a result, total mastectomy is the most common form of surgical treatment. The lack of radiation therapy is also reflected in the recommendation by local stakeholders to exclude reporting on extracapsular extension of metastatic carcinoma in lymph nodes.

None of the pathology reports had all data elements categorized as essential by local stakeholders, while, depending on the site, only 9% to 65% of cases contained all 5 parameters recommended by BCI 2.5 for basic resource settings (ie, tumor size, lymph node status, histologic type, histologic grade, and TNM stage). These findings are commensurate with other published reports on the completeness of breast pathology reports in LMICs. A few studies in Nigeria also found a significant rate of missing critical data elements in their breast pathology reports, including histologic tumor type (8%–72%), tumor size (0%–50%), tumor grade (10%–60%), margin status (36%–52%), lymph node status (60%–71%), and LVI status (86%–97%), but which varied depending on the time period analyzed and institution.^17–19  Another study from South Africa found that 26% (298 of 1158) of breast cancer surgical specimen reports were incomplete, with tumor grade as the most common missing parameter.²⁰  A review of breast cancer pathology reports in Ethiopia found that only 62% (256 of 417) of reports included tumor size, histologic type, histologic grade, and lymph node status.²¹  Incomplete reports can have profoundly negative consequences on patient management and potentially outcomes. For example, critical data elements, such as tumor size and the number of involved lymph nodes, are essential for staging, prognosis, and identifying optimal treatment, while ER and HER2 testing results are required for oncologists to determine whether patients are eligible for hormonal or anti-HER2 therapy. Surgical margin status also dictates whether or not patients need additional surgery to avoid early recurrence. Delays in treatment can also occur if data elements are missing and oncologists or surgeons need to contact the laboratory to obtain the information. Trust between physicians is also jeopardized when pathology reports are incomplete, affecting the ability to communicate effectively and provide multidisciplinary care. Therefore, there is an urgent need to improve pathology reporting practices in SSA.

Synoptic reporting has been shown to increase the completeness of reports across organ sites and a range of resource settings in comparison to narrative reporting.^22–25  Completeness rates can be greater than 95% when data elements are well formatted, optional elements are excluded, and electronic reporting tools are used.^3,26,27  While synoptic reporting initially appears to increase the time it takes to generate a report, successful implementation typically results in a reduction of time spent on generating a report.^4,22  Moreover, a study from the Netherlands found that an increase in the use of synoptic reporting from 12% to 79% of cases resulted in more frequent inclusion of HER2 results, and in turn, increased administration of anti-HER2 therapy to eligible patients.²⁴  This finding demonstrates that improving the quality of pathology reports can have direct impact on clinical care.

Invasive breast cancer reporting guidelines, and their associated synoptic templates, have been published by several professional pathology organizations, including the CAP, ICCR, and RCP, all of which include similar recommendations for core (essential) and noncore (optional) data.^7–9,28,29  The synoptic templates published by these organizations contain numerous optional elements and were developed with limited input from stakeholders who practice in LMICs, impeding their implementation in low-resource settings.²⁹  Although these templates are readily available online, they are typically only accessible in a static format, such as a Portable Document Format (pdf) or Microsoft Word document, restricting their dissemination and the ability to readily customize them. In addition, while the CAP offers an electronic product that can be licensed, the cost is prohibitively expensive for LMIC laboratories.³⁰ 

Although synoptic reporting has been adopted widely in high-income settings,^22,24,31  few studies have described their implementation in low- and middle-income countries. Publications from Malaysia, Nigeria, and Nicaragua describe the successful use of synoptic templates but do not describe how the templates were implemented.^23,25,32  It is therefore unclear what are the most significant facilitators and barriers to implementing synoptic reporting in LMICs. A survey across 27 hospitals in Canada reported that difficulties in implementing synoptic reporting were primarily technology-related factors.⁴  Despite those challenges, pathologists preferred synoptic reports over narrative reports because they felt that templates reduced the chance for error or forgetting to include a specific parameter that is significant for a cancer case.⁴  Although the survey evaluated implementation of synoptic reporting in a high-resource setting, it is highly likely that technology-related factors will also be a major factor in LMICs. Other potential barriers to successful implementation include the lack of subspecialty training among pathologists, the need to maintain and update the templates, and pathologists who may be resistant to changes in practice.

Several limitations of this study should be noted. Although distinct geographic regions in SSA were included in this study, only 3 hospitals were evaluated, all of which were large referral hospitals. However, the aim of this study was not to describe the state of breast cancer pathology services across SSA but to conduct a situational analysis that would allow us to create a roadmap for how to improve breast cancer reporting at select hospitals that provide care to a large proportion of the nation’s patients with cancer, and with which the authors already have preexisting partnerships. Interestingly, the challenges to providing high-quality diagnostic services were similar across all 3 sites. Another limitation is that only a small subset of pathology reports was reviewed. In Mozambique, reports from only 3 malignant core needle biopsies or incisional surgical biopsies were made available for review. The limited number of biopsies is because most primary diagnoses of breast masses are made by fine-needle aspiration biopsy, which is then typically followed by total mastectomy with or without neoadjuvant chemotherapy. Although the pathology report review was concise, it was sufficient for the purposes of the needs assessment, and our intention was not to conduct a thorough audit. A final limitation is that while the synoptic template data elements were reviewed by a set of multidisciplinary stakeholders at least twice, the authors were unable to have all stakeholders meet to review the template at the same time.

In conclusion, this study identified urgent gaps in breast pathology services at 3 large referral hospitals in distinct geographic regions in SSA. Using the data from the needs assessment, a locally adapted synoptic tumor template was designed from already published invasive breast cancer reporting guidelines. For the next phase of our initiative, we intend to implement an open access interactive web-based version of the synoptic templates, which in several studies has been shown to ensure higher rates of completeness.^26,33–35  This needs assessment will inform the implementation plan to introduce synoptic reporting at the 3 large referral hospitals included in this study and will potentially make durable improvements in breast pathology services in other SSA countries and beyond.