Gross Evaluation of Breast Carcinomas Post–Neoadjuvant Chemotherapy Without Radio-opaque Clip Insertion: An Alternative Option Open Access

The objective of the study was to evaluate the number of sections to be sampled serially in a post-NACT specimen without prior placement of radio-opaque clips and to compare the response and residual tumor by residual cancer burden (RCB)⁵  method in the single largest cross-sectional area and entire tumor bed.

This was a single-center study approved by the Institutional Review Board. In this study, 50 breast carcinoma cases post-NACT were grossed by a single pathologist prospectively. Before grossing the prechemotherapy and postchemotherapy mammograms were reviewed to understand the location of tumor. Specimens were serially sliced at 5-mm intervals and routinely fixed and processed. The tumor bed was identified, and 1 slice was selected as most representative of the entire tumor bed (sample I). Subsequently the entire tumor bed was submitted in grids of multiple slices (sample II). An image of the sliced specimen in the form of a photograph or drawing was taken and then used as a map for the sections submitted, so that the microscopic findings of residual disease in the breast could be easily comprehended (Figure 1, A through F). The average number of cassettes required to submit the tumor bed per case was calculated for both methods. Clinicopathologic parameters were studied and cases were divided based on their biomarker status into 4 groups as hormone receptor (HR)–positive (estrogen receptor (ER)⁺/progesterone receptor (PR)⁺/HER2⁻), triple positive (ER⁺/PR⁺/HER2⁺), HER2 enriched (ER⁻/PR⁻/HER2⁺), and TNBC (ER⁻/PR⁻/HER2⁻). pCR was defined as having no residual tumor in breast and lymph nodes.³  RCB scoring for both groups was compared using a χ² test. P < .05 was considered significant, and κ values were calculated for sample I as compared to sample II (gold standard).

Median patient age was 45.5 years, and median clinical tumor size was 5.5 cm (n = 50). Thirty-three of 50 (66%) patients had locally advanced breast carcinoma whereas 17 of 50 (34%) patients had operable breast cancer. Twenty of 50 patients (40%) underwent BCS whereas 30 of 50 (60%) underwent mastectomy. Twelve of 50 patients were HR positive (24%), 11 of 50 (22%) were triple positive, 15 of 50 were HER2 positive (30%), and 12 of 50 were classified as TNBC (24%). The clinicopathologic details are summarized in Table 1.

A mean of 8 blocks per case were made for a single slice with the largest cross-sectional area (sample I) whereas an average of 26 blocks were made for sampling the entire tumor bed (sample II). By sample I, 23 cases had complete response in the breast whereas by sample II, 21 cases had complete response in breast. Two cases, labeled by sample I as having no residual tumor, were classified as RCB I by sample II. Both were HER2 receptor positive, one of which was hormone receptor (ER, PR) positive as well. Both the cases showed singly scattered isolated tumor cells in the stroma and comprised less than 5% of the tumor bed (Figure 2, A through D). The residual tumor was found in the slice adjacent to the selected slice.

As all firm areas (tumor bed) were grossed in sample II, calculation of RCB was accurate and considered the gold standard. The RCB was also calculated in sample I and compared with that in sample II. The RCB classes as per both the methods are displayed in Table 2. The RCB scores as measured by the 2 methods are demonstrated in Figure 3.

Distribution of cases according to their biomarker status and RCB class is depicted in Table 3.

The sensitivity of sample I to detect complete response was 100%, specificity was 93.1%, and accuracy was 96%. Significant correlation of RCB scoring was seen between the 2 methods using the χ² test (P < .001). With respect to RCB scoring, there was 94% concordance between the 2 methods. The κ value for sample I was 0.915, which showed very good correlation with sample II.

Margins were involved by invasive tumor in 1 case (HER2 enriched) and involved by tumor bed in 13 cases (5: HER2 enriched, 4: TNBC, 3: triple positive, 1: HR positive). Margins of 36 cases were free of tumor and tumor bed.

Tumor shrinkage pattern was assessed in all cases with residual tumor in the breast (n = 29). Out of the 29 cases, 5 cases showed concentric shrinkage, 14 cases showed singly scattered tumor cells, and 6 cases showed multiple scattered tumor foci. Four cases had largely residual viable tumor and no specific pattern was noted.

Correlation of biomarker status with pattern of shrinkage is summarized in Table 4.

The median event-free survival (EFS) for 50 cases was 21 months. No recurrences were seen in cases with concentric shrinkage (2-year EFS: 100%). Two-year EFS of cases with singly scattered tumor cells was 53.5%, whereas for those with multiple nodular tumor foci EFS was 75%, and with complete response it was 100% (P = .005). The case with an invasive tumor involving margins had recurrence within 2 years. There was no significant difference in EFS between cases with tumor bed involving margins and those with the tumor bed free of margins. Also, RCB classes showed no significant difference in EFS; however, this was most likely due to a short follow-up period.

Quantification of tumor response is essential since it forms an important prognostic factor to predict survival.⁹  To classify this tumor response, many grading systems have been proposed and have evolved over the years including the Miller-Payne system,¹⁰  Residual Disease in Breast and Nodes,¹¹  the Chevalier method,¹²  the Sataloff method,¹³  and RCB, to name a few.^5,14  Different systems have different definitions of pCR¹⁵  and take different parameters into consideration (tumor cellularity, presence of in situ carcinoma, percentage of response, histologic grade, etc). The characterization of post-NACT response differs from country to country. The RCB system by the MD Anderson Cancer Center is the most clinically validated and widely used system.^14,16  It has been included as a core element in the International Collaboration on Cancer Reporting (ICCR) dataset.¹⁷ 

RCB classification is based on a scoring system that uses a web-based calculator¹⁴  and includes microscopic tumor bed size, percentage of cellularity of invasive and in situ components, number of positive axillary lymph nodes, and the largest dimension of metastatic deposit.⁵  Thus, appropriate sampling of the tumor bed is of paramount importance for accurate RCB scoring. Incomplete tumor bed gross evaluation would cause underestimation of residual tumor and thus an incorrect RCB score. RCB class has been proven to be an independent prognostic factor in post-NACT breast carcinoma.^9,16 

Several guidelines discuss gross evaluation of post-NACT specimens.¹⁸  While grossing a post-NACT breast carcinoma specimen, the size of the primary tumor bed should be measured, and macroscopically visible lesions should be described, mapped, and sampled extensively.^18–22  BCS or a wide local excision should be grossed systematically by inking the specimen according to orientation, serially slicing it at 5 mm intervals, and submitting it entirely (if <5 cm in greatest dimension or <30 g) in a way such that reconstruction of the specimen is possible at the time of microscopic evaluation with the help of a diagram.^18,23  A tumor bed extending to the margins should be documented.^18,24  In case of a large lumpectomy or mastectomy, the specimen should be serially sliced at 5–10-mm intervals and the tumor bed should be initially sampled by submitting 5 blocks at every 1–2 cm of the tumor bed, with a maximum of 25 blocks.^24,25  The US Food and Drug Administration recommends a minimum of 1 block per centimeter of pretreatment tumor size or at least 10 blocks, whichever is greater.^19,24,25 

All of the grossing techniques described worldwide prefer localization of tumor bed by radio-opaque clip. This radio-opaque clip in the tumor is placed before or during the administration of NACT. They are especially useful in cases in which there is complete or near-complete response, which would make it difficult for surgeons to localize the tumor bed. The placement of radio-opaque clips in patients undergoing BCS is associated with better local control.²⁶  Pre-NACT clip placement in the axillary lymph node suspected of being metastatic and localization and removal of the same post-NACT is known to improve the accuracy of sentinel lymph node biopsy in post-NACT tumors.²⁷  Clip placement is not a routine practice due to the cost considerations and thus clinical examination and radiography are relied on for tumor bed identification. We have shown that as long as radiologic response is taken into account and correlated with the gross examination, pCR calculations are 90% accurate and calculations of RCB are also feasible using this technique. Nevertheless, there is a need for a wider multidisciplinary team involved in breast care to agree and discuss the allocation of resources, according to clinical settings. However, it is difficult to have availability of resources and experts in every field in the peripheral resource-limited setting.

It is very important to know the biomarker status of the tumor before gross evaluation of the resection specimen. HER2 receptor–positive tumors have a characteristic pattern of residual disease in the form of single scattered, isolated cells in a hyalinized vascular stroma. This response may be missed with our sample I method, hence, it is advisable to sample extra slices beyond the single-slice method to avoid mislabeling the specimen as pCR. Though not seen in our study it has been noted that TNBCs also have a high rate of pCR and can also have residual tumor in the form of single scattered tumor cells. Thus, additional sectioning from adjacent slices is warranted in HER2-enriched and TNBC tumors with no grossly residual tumor. Cases with this pattern of shrinkage are also seen to have earlier recurrence as compared to the other patterns.

Sample I is a more cost-effective sampling method, even in mastectomy. The use of sample I technique resulted in a process that was 33% of the cost of sample II, thus decreasing the expense to a marked extent. In resource-constrained settings, this would be a good method where tumor is not previously localized by radio-opaque clips. Thus, implementation of this method would decrease costs without compromising the quality of care.

The study has a few limitations. The sample size of the cohort is small given the laborious task of sampling the whole tumor bed, and additional validation is recommended in a larger cohort of breast carcinoma specimens of different receptor profiles. We purposefully ensured that the gross evaluation was done by a single pathologist to ensure uniformity of expertise for this pilot exercise. Further study of this method with different pathologists at different centers is essential before extrapolating the results and considering implementation in the general population.

We thus conclude that single-slice method of grossing with information on mammographic findings is useful to calculate RCB even if tumor is not localized by radio-opaque clip/wire prior to NACT, except in a few HER2-positive tumors.