This diagnostic study assesses the accuracy of post–neoadjuvant chemotherapy image-guided biopsy to predict residual cancer in the breast.
Key Points
Question
Can image-guided breast biopsy predict the presence of residual cancer in patients treated with neoadjuvant chemotherapy?
Findings
In this diagnostic study of 166 women, the use of a standardized protocol involving image-guided vacuum-assisted biopsy to retrieve 6 or more representative samples of a tumor bed measuring 2 cm or smaller was found to reliably identify the presence of residual cancer and predict pathologic complete response with a false-negative rate less than 5%.
Meaning
Post–neoadjuvant chemotherapy image-guided biopsy may identify pathologic complete response in the breast in selected patients; these findings can inform the design of de-escalation trials testing the elimination of surgery in exceptional responders.
Abstract
Importance
Image-guided breast biopsy of a residual imaging abnormality or tumor bed after neoadjuvant chemotherapy (NACT) is increasingly used to assess residual cancer, facilitate risk-adaptive surgery, and potentially identify exceptional responders in whom local therapy may be de-escalated.
Objective
To further assess the accuracy of post-NACT image-guided biopsy to predict residual cancer in the breast.
Design, Setting, and Participants
This diagnostic study analyzed multicenter patient-level data of patients with breast cancer who were treated with NACT and underwent image-guided biopsy before surgery at Royal Marsden Hospital in London, UK; Seoul National University Hospital in Seoul, South Korea; and MD Anderson Cancer Center in Houston, Texas. Data were analyzed from June to July 2019.
Main Outcomes and Measures
Diagnostic accuracy of post-NACT image-guided biopsy. Final surgical pathology was used as reference standard.
Results
Data from 166 women were analyzed. The median (range) age was 49 (25-76) years. The median (range) tumor size on pretreatment and posttreatment imaging was 33.5 (12-100) mm and 10 (0-100) mm, respectively. The overall pathologic complete response rate was 51.2% (n = 85) (16.1% [5 of 31] for hormone receptor–positive/ERBB2 (formerly HER2)-negative; 44.7% [21 of 47] for hormone receptor–positive/ERBB2-positive; 69% [20 of 29] for hormone receptor–negative/ERBB2-positive; and 66.1% [39 of 59] for triple negative). The majority (143 [86.1%]) underwent image-guided vacuum-assisted biopsy (VAB), and 23 had core-cut biopsy. The median (range) needle gauge was 10 (7-14), and the median (range) number of samples was 6 (2-18). When image-guided biopsy (VAB and core-cut biopsy) was representative (159 [95.8%]), the false-negative rate across the whole cohort was 18.7% (95% CI, 10.6%-29.3%). Subgroup analysis of patients with a complete/partial clinical response and residual imaging abnormality of 2 cm or smaller with at least 6 VABs taken (76 [45.8%]) demonstrated a false-negative rate of 3.2% (95% CI, 0.1%-16.7%), a negative predictive value of 97.4% (95% CI, 86.5%-99.9%), and an overall accuracy of 89.5% (95% CI, 80.3%-95.3%).
Conclusions and Relevance
This large multicenter pooled data analysis suggests that a standardized protocol using image-guided VAB of a tumor bed measuring 2 cm or smaller with 6 or more representative samples allows reliable prediction of residual disease. These results could inform the design of de-escalation trials in NACT exceptional responders testing the safety of eliminating surgery.
Introduction
The use of neoadjuvant chemotherapy (NACT) in early-stage breast cancer is increasing.1 Its use is likely to further increase as results from clinical trials support its prognostic role demonstrating potential benefit for continuing systemic therapy in the adjuvant setting in patients with residual disease after NACT.2,3 Advances in systemic therapies and the introduction of targeted therapies have resulted in improved pathologic complete response (pCR) rates, particularly in triple-negative (TN) and ERBB2 (formerly HER2)-positive breast cancers. In these subtypes, pCR rates exceeding 60% have been reported,4,5,6 suggesting that in selected patients, there might not be benefit from surgery, especially with the routine use of adjuvant radiotherapy and systemic therapy.
In the modern era of precision medicine and personalized oncology, there is a continuous effort to tailor treatments based on tumor biology and response to systemic therapies. This includes a trend toward de-escalation of locoregional therapy including surgery and/or radiation in the breast and the axilla after NACT, allowing less radical operative procedures and radiation therapy.7,8,9,10 However, despite advances in medical imaging, there is no single method or combination of modalities that can accurately predict pCR in the breast preoperatively and thus reliably identify exceptional responders who could be spared further local treatments including surgical intervention.11 Image-guided percutaneous biopsy of the residual imaging abnormality/tumor bed after completion of NACT is gaining popularity as a predictive tool, especially in selected patients, and published data have shown promising results.12,13,14,15,16
The primary aim of this study was to further assess the accuracy of post-NACT image-guided biopsy to identify residual cancer in the breast by performing a pooled analysis of individual patient data from 3 institutions (Royal Marsden Hospital in London, UK; Seoul National University Hospital in Seoul, South Korea; and MD Anderson Cancer Center in Houston, Texas) practicing this technique. An additional aim was to identify any variables (patient demographics, tumor characteristics, imaging features, or biopsy technique) that might be associated with the observed accuracy (eg, false-negative rate [FNR], sensitivity, specificity, negative predictive value [NPV], and overall accuracy) of post-NACT image-guided biopsy. The findings could be helpful in defining an optimal assessment protocol and assist in the design of de-escalation trials in selected groups of exceptional responders to NACT testing the safety of omission of surgery and/or radiotherapy.
Methods
Study Design
This is a multicenter pooled analysis of individual patient data. Following institutional review board approval from each participating center (Royal Marsden Hospital in London, UK; Seoul National University Hospital in Seoul, South Korea; and MD Anderson Cancer Center in Houston, Texas), predetermined anonymized data were collected. Recruitment of patients and post-NACT image-guided biopsy technique in each participating institution has been previously described.14,15,16,17 Eligible patients for the pooled analysis were women older than 18 years with diagnosis of invasive breast cancer of any subtype that were treated with NACT per local guidelines, had at least partial clinical/imaging response, and underwent image-guided biopsy of the residual imaging breast abnormality before proceeding to breast surgery. Women with no clearly visible residual tumor or tumor microcalcifications in the absence of a marker clip or those who did not have breast surgery were excluded (eAppendix in the Supplement). Imaging complete response was defined as the absence of imaging abnormality on mammography, ultrasonography, or magnetic resonance imaging (MRI) (if available). Patients underwent image-guided biopsy to sample the residual imaging breast abnormality. If there was no residual abnormality, the marker clip representing the tumor bed was sampled after confirming that there was no clip migration. The cores were radiographed to confirm retrieval of the marker clip and the presence of targeted microcalcifications, if applicable. After image-guided biopsy, a new marker clip (or a guide wire if the procedure was performed on the day of surgery) was placed to guide subsequent surgery. If after sampling of the target area, the original marker clip was not retrieved for technical reasons, this was recorded.
Final surgical pathology was used as the reference standard. Pathologic complete response was defined as no residual disease in the breast (invasive and ductal carcinoma in situ, ypT0) for the purposes of this analysis. In case of breast pCR, the tumor bed was characterized by the presence of edematous stroma with inflammatory cell and macrophage infiltration and stromal fibrosis. A biopsy was considered nonrepresentative if the target area was not appropriately sampled (radiograph not showing the marker clip or residual microcalcifications as applicable), or if histopathology did not describe cancer or tumor bed defined as presence of hemorrhage with granulation tissue and early scar formation.
The primary end point was determination of the FNR of post-NACT image-guided biopsy in the pooled cohort. The FNR was defined as no residual cancer in the image-guided biopsy (index test = negative) but presence of residual cancer as reported by histopathology of the surgical specimen (reference test = positive). Secondary end points included overall accuracy, sensitivity, specificity, and NPV.
A secondary aim was to perform a subgroup analysis to investigate potential associations between tumor, imaging, and biopsy technique–related variables and the diagnostic accuracy of post-NACT image-guided biopsy and to try to define an optimal assessment protocol.
Statistical Analysis
Simple descriptive statistics were used. Continuous variables were summarized using median and range, while categorical variables were summarized by frequency and corresponding percentages. False-negative rate, sensitivity, specificity, NPV, and overall accuracy were calculated and reported with 95% CIs using the binomial exact method. The proportion of concordant cases between image-guided biopsy (index test) and surgical histopathology (reference test) was compared between centers using χ2 test for independence. Univariate logistic regression analysis was used to explore associations between various patient, tumor, image, and biopsy technique characteristics and the concordance between image-guided biopsy and final surgical histopathology. Regression coefficients were tested for significance using 2-sided Wald tests with significance threshold of .05. Statistical analyses were performed using SPSS version 23.0 (IBM Corp) and Stata version 13.1 (StataCorp). Data were analyzed from June to July 2019.
Results
A total of 166 eligible patients were included in the overall analysis. A flow diagram of study participants is presented in the Figure. Patient demographics and tumor characteristics at diagnosis are presented in Table 1.
Figure. Flow Diagram of Patients in the Pooled Analysis.
NACT indicates neoadjuvant chemotherapy; VAB, vacuum-assisted biopsy.
Table 1. Patient Demographics and Tumor Characteristics in 166 Patients.
| Characteristic | No./total No. (%) |
|---|---|
| Age, median (range), y | 49 (25-76) |
| Histologic subtype | |
| Invasive ductal carcinoma | 160/166 (96.4) |
| Invasive lobular carcinoma | 6/166 (3.6) |
| Histologic grade | |
| 1 | 0/166 (0) |
| 2 | 42/166 (25.3) |
| 3 | 124/166 (74.7) |
| Tumor subtype | |
| HR+/ERBB2– | 31/166 (18.7) |
| HR+/ERBB2+ | 47/166 (28.3) |
| HR–/ERBB2+ | 29/166 (17.5) |
| HR–/ERBB2– | 59/166 (35.5) |
| Maximum tumor size pre-NACT, median (range), mm | 33.5 (12-100) |
| Tumor cT stage (TNM) pre-NACT, No. (%) | |
| T1 | 21 (13) |
| T2 | 117 (70) |
| T3 | 28 (17) |
| Multifocal carcinoma, No. (%) | 37 (22) |
| Multicentric carcinoma, No. (%) | 2 (1) |
| Microcalcifications, No. (%) | |
| Suspicious microcalcifications | 59/151 (39.1) |
| Microcalcifications extending beyond index tumor | 32/151 (21.2) |
| Maximum tumor size post-NACT, median (range), mm | 10 (0-100) |
| Tumor ycT stage (TNM) post-NACT (%) | |
| T1 | 142 (85.5) |
| T2 | 22 (13.3) |
| T3 | 2 (1.2) |
| DCIS component present on initial diagnostic biopsy (n = 123) | 35/123 (28.5) |
| Pathologic complete response in breast | |
| All subtypes | 85/166 (51.2) |
| HR+/ERBB2– | 5/31 (16.1) |
| HR+/ERBB2+ | 21/47 (44.7) |
| HR–/ERBB2+ | 20/29 (69) |
| HR–/ERBB2– | 39/59 (66.1) |
Abbreviations: DCIS, ductal carcinoma in situ; cT, clinical pretreatment T; HR, hormone receptor; NACT, neoadjuvant chemotherapy; TNM, tumor, node, metastasis; ycT, clinical posttreatment T stage.
Imaging Findings
The median (range) maximum tumor size on pre-NACT imaging (mammogram, ultrasonography, or MRI when performed) was 33.5 (12-100) mm. Multifocal disease was present in 37 patients (22.3%), and 2 (1.2%) had multicentric disease. Data on the presence of microcalcifications were available in 151 patients; these were reported as suspicious in 59 (39.1%) and extending beyond the index tumor in 32 (21.2%). The median (range) maximum size of the residual breast abnormality on post-NACT imaging (mammogram, ultrasonography, or MRI when performed) was 10 (0-100) mm.
NACT Regimens
All patients received anthracycline- and/or taxane-based chemotherapy per local and international guidelines. Of 76 individuals with ERBB2-positive breast cancer, in addition to chemotherapy, 12 (15.8%) received single anti-ERBB2 therapy with trastuzumab, and the rest (64 [84.2%]) received dual anti-ERBB2 treatment with trastuzumab and pertuzumab.
Image-Guided Biopsy
After completion of NACT, all 166 patients underwent image-guided biopsy of the residual imaging abnormality or tumor bed. In 6 patients (3.6%), no marker clip was inserted in the breast, but they all had clearly visible residual tumor and a representative image-guided biopsy. A total of 143 patients (86.1%) underwent vacuum-assisted biopsy (VAB) and 23 (13.9%) had core-cut biopsy. The biopsy was performed under ultrasonography or stereotactic guidance in 129 patients (77.7%) and 37 patients (22.3%), respectively. The median (range) needle gauge was 10 (7-14), and the median (range) number of samples taken was 6 (2-18). The biopsy was found to be representative in 159 patients (95.8%). Nonrepresentative biopsies were excluded from further analysis.
Histopathology Results
The overall pCR rate was 51.2% (n = 85). The pCR rates according to cancer subtype are presented in Table 1. Seventy patients (42%) had biopsy-proven metastatic axillary lymph nodes at diagnosis. Of these, 45 (64.3%) achieved nodal pCR after NACT. Among 96 clinically node-negative patients, 2 of 56 (3.6%) who had breast pCR were found to have disease in the axillary lymph nodes on surgical histopathology. In the clinically node-positive group (n = 70), 3 of 29 patients (10.3%) who had breast pCR had residual disease in the axilla.
Overall Analysis
The FNR of image-guided biopsy in the whole cohort was 18.7% (95% CI, 10.6%-29.3%) with an NPV of 84.3% (95% CI, 75%-91.1%) and overall accuracy of 85.5% (95% CI, 79.1%-90.6%). Final histopathology details of patients with an image-guided biopsy showing no residual cancer but presence of residual cancer in the surgical specimen are presented in the eTable in the Supplement. Details on sensitivity and specificity and results following subgroup analysis based on breast cancer subtype are presented in Table 2.
Table 2. Overall Cohort Diagnostic Accuracy of Post-NACT Image-Guided Biopsy.
| Characteristic | % (95% CI) |
|---|---|
| All histologic types (n = 159)a | |
| Sensitivity | 81.3 (70.7-89.4) |
| Specificity | 89.3 (80.6-95.0) |
| False-negative rate | 18.7 (10.6-29.3) |
| Negative predictive value | 84.3 (75-91.1) |
| Overall accuracy | 85.5 (79.1-90.6) |
| Invasive ductal carcinoma | |
| All subtypes (n = 153) | |
| Sensitivity | 82.6 (71.6-90.7) |
| Specificity | 89.3 (80.6-95) |
| False-negative rate | 17.4 (9.3-28.4) |
| Negative predictive value | 86.2 (77.1-92.7) |
| Overall accuracy | 86.3 (79.8-91.3) |
| HR+/ERBB2– (n = 25) | |
| Sensitivity | 90 (68.3-98.8) |
| Specificity | 100 (47.8-100)b |
| False-negative rate | 10 (1.2-31.7) |
| Negative predictive value | 71.4 (29-96.3) |
| Overall accuracy | 92 (74-99) |
| HR+/ERBB2+ (n = 40) | |
| Sensitivity | 85 (62.1-96.8) |
| Specificity | 85 (62.1-96.8) |
| False-negative rate | 15 (3.2-37.9) |
| Negative predictive value | 85 (62.1-96.8) |
| Overall accuracy | 85 (70.2-94.3) |
| HR–/ERBB2+ (n = 29) | |
| Sensitivity | 66.7 (29.9-92.5) |
| Specificity | 95 (75.1-99.9) |
| False-negative rate | 33.3 (7.5-70.1) |
| Negative predictive value | 86.4 (65.1-97.1) |
| Overall accuracy | 86.2 (68.3-96.1) |
| HR–/ERBB2– (n = 59) | |
| Sensitivity | 80 (56.3-94.3) |
| Specificity | 87.2 (72.6-95.7) |
| False-negative rate | 20 (5.7-43.7) |
| Negative predictive value | 89.5 (75.2-97.1) |
| Overall accuracy | 84.7 (73-92.8) |
Abbreviations: HR, hormone receptor; NACT, neoadjuvant chemotherapy.
Patients with nonrepresentative image-guided biopsy (n = 7) were excluded from the analysis.
One-sided 97.5% CI.
Univariate logistic regression analysis did not show any association between patient, tumor, imaging, or biopsy technique–associated variables and the concordance between the image-guided biopsy and surgical histopathology. The only exception was a nonrepresentative biopsy (odds ratio, 0.064; 95% CI, 0.12-0.352; P = .002) (Table 3).
Table 3. Univariate Logistic Regression of Patient, Tumor, Imaging, and Biopsy Technique-Related Variables and Concordance Between Imaging-Guided Biopsy and Surgical Histopathology.
| Variable | Odds ratio (95% CI) | P value |
|---|---|---|
| Age | 0.989 (0.954-1.024) | .53 |
| Invasive ductal carcinoma | ||
| No | 0.276 (0.044-1.735) | .17 |
| Yes | 1 [Reference] | |
| Invasive lobular carcinoma | ||
| No | 2.700 (0.469-15.541) | .27 |
| Yes | 1 [Reference] | |
| Histologic grade | ||
| 2 | 1.223 (0.458-3.271) | .69 |
| 3 | 1 [Reference] | |
| HR+/ERBB2– | ||
| No | 0.988 (0.342-2.852) | .98 |
| Yes | 1 [Reference] | |
| HR+/ERBB2+ | ||
| No | 1.329 (0.550-3.213) | .53 |
| Yes | 1 [Reference] | |
| HR–/ERBB2+ | ||
| No | 0.793 (0.252-2.496) | .69 |
| Yes | 1 [Reference] | |
| HR–/ERBB2– | ||
| No | 0.890 (0.372-2.128) | .79 |
| Yes | 1 [Reference] | |
| DCIS component on initial diagnostic biopsy | ||
| No | 2.423 (0.903-6.5) | .08 |
| Yes | 1 [Reference] | |
| Maximum tumor size pre-NACT | 0.999 (0.977-1.022) | .95 |
| Pre-NACT cT stage (TNM) | ||
| 1 | 1 [Reference] | .56 |
| 2 | 1.838 (0.595-5.680) | .29 |
| 3 | 1.438 (0.357-5.795) | .61 |
| Multifocal disease | ||
| No | 0.753 (0.264-2.148) | .60 |
| Yes | 1 [Reference] | |
| Associated microcalcifications | ||
| No | 2.081 (0.863-5.021) | .10 |
| Yes | 1 [Reference] | |
| Microcalcifications extending beyond index tumor | ||
| No | 2.146 (0.823-5.593) | .12 |
| Yes | 1 [Reference] | |
| Maximum tumor size post-NACT | 0.984 (0.952-1.016) | .32 |
| Post-NACT ycT stage (TNM) | ||
| 1 | 1 [Reference] | .15 |
| 2 | 0.149 (0.163-1.318) | |
| 3a | NA | NA |
| Type of image-guided biopsy | ||
| Vacuum assisted | 0.744 (0.205-2.703) | .65 |
| Core cut | 1 [Reference] | |
| Size of biopsy needle | 1.105 (0.874-1.396) | .41 |
| No. of biopsy samples | 0.966 (0.836-1.115) | .63 |
| Method of image-guidance | ||
| Ultrasonography | 1.597 (0.635-4.015) | .32 |
| Stereotactic | 1 [Reference] | |
| Biopsy representative | ||
| No | 0.064 (0.12-0.352) | .002 |
| Yes | 1 [Reference] |
Abbreviations: cT, clinical pretreatment T; DCIS, ductal carcinoma in situ; HR, hormone receptor; NA, not applicable; NACT, neoadjuvant chemotherapy; TNM, tumor, node, metastasis; ycT, clinical posttreatment T.
Only 2 patients had ycT T3 stage tumors, and both had concordant findings between image-guided biopsy and surgical histopathology. Therefore, analysis was not feasible.
The proportion of concordant cases between image-guided biopsy and surgical histopathology result was not found to be significantly different across each participating center on the χ2 test for independence (χ2 = 1.543; P = .46).
Subgroup Analysis
Further exploratory subgroup analysis was performed to investigate if an optimal assessment protocol could be determined, as defined by an FNR conventionally set as less than 5%. Patients with invasive lobular carcinoma (n = 6) were excluded from the subgroup analysis because of the small number of cases and the inherent differences in growth pattern and potential challenges in determining disease extent on imaging. A maximum size of residual imaging abnormality or tumor bed of up to 2 cm was selected as this is the cutoff for T1 tumors based on the TNM (tumor, node, metastasis) staging system and was regarded by the participating radiologists as an optimal size that would allow adequate sampling in patients with near-complete imaging response. The use of image-guided VAB was selected because of the ability of the technique to yield more tissue for histopathologic examination and ease of sampling.18 A minimum number of 6 representative core samples was chosen because this was the median number of samples in this pooled cohort. Following application of these criteria (n = 98) and exclusion of patients with missing values (number of biopsy samples, n = 22), 76 patients were included in the final exploratory analysis (Figure). Of these, 26 patients (34.2%) had microcalcifications, which in 15 cases were extending beyond the tumor. The FNR was found to be 3.2% (95% CI, 0.1%-16.7%). Only 1 patient with hormone receptor–positive/ERBB2-positive invasive ductal carcinoma, with a VAB showing no evidence of cancer, had residual cancer in final surgical histopathology. In this case, the residual cancer was less than 1 mm invasive ductal carcinoma associated with ductal carcinoma in situ with a cellularity of less than 1% representing minimal residual tumor burden. Results of diagnostic accuracy of image-guided VAB across this subgroup and based on breast cancer subtype are presented in Table 4. There was no statistically significant difference in the proportion of concordant cases across the 3 centers (χ2 = 3.341; P = .19) in the subgroup analysis.
Table 4. Diagnostic Accuracy of Post–Neoadjuvant Chemotherapy Image-Guided Vacuum-Assisted Biopsy in Patients With Residual Imaging Abnormality/Tumor Bed ≤2 cm and at Least 6 Representative Samples.
| Characteristic | % (95% CI) |
|---|---|
| All subtypes (n = 76) | |
| Sensitivity | 96.8 (83.3-99.9) |
| Specificity | 84.4 (70.5-93.5) |
| False-negative rate | 3.2 (0.1-16.7) |
| Negative predictive value | 97.4 (86.5-99.9) |
| Overall accuracy | 89.5 (80.3-95.3) |
| HR+/ERBB2– (n = 10) | |
| Sensitivity | 100 (59-100)a |
| Specificity | 100 (29.2-100)a |
| False-negative rate | 0 (0-40.1)a |
| Negative predictive value | 100 (29.2-100)a |
| Overall accuracy | 100 (69.1-100)a |
| HR+/ERBB2+ (n = 21) | |
| Sensitivity | 90 (55.5-99.7) |
| Specificity | 72.7 (39-94) |
| False-negative rate | 10 (0.3-44.5) |
| Negative predictive value | 88.9 (51.7-99.7) |
| Overall accuracy | 80.9 (58.1-94.5) |
| HR–/ERBB2+ (n = 11) | |
| Sensitivity | 100 (29.2-100)a |
| Specificity | 100 (63.1-100)a |
| False-negative rate | 0 (0-70.8)a |
| Negative predictive value | 100 (63.1-100)a |
| Overall accuracy | 100 (71.5-100)a |
| HR–/ERBB2– (n = 34) | |
| Sensitivity | 100 (71.5-100)a |
| Specificity | 82.6 (61.2-95) |
| False-negative rate | 0 (0-28.5)a |
| Negative predictive value | 100 (82.3-100)a |
| Overall accuracy | 88.2 (72.5-97.7) |
| ERBB2+ and HR–/ERBB2– (n = 66) | |
| Sensitivity | 95.8 (78.9-99.9) |
| Specificity | 83.3 (68.6-93) |
| False-negative rate | 4.2 (0.1-21.1) |
| Negative predictive value | 97.2 (85.5-99.9) |
| Overall accuracy | 87.9 (77.5-94.6) |
Abbreviation: HR, hormone receptor.
One-sided 97.5% CI.
Discussion
The results of this multicenter pooled analysis of patient-level data showed that image-guided biopsy at completion of NACT is associated with an FNR of 18.7% across the whole cohort (FNR, 17.4% for patients with invasive ductal cardinoma; range, 10%-33% depending on breast cancer subtype) (Table 2). These findings are similar to previously published data in unselected patients who underwent image-guided biopsy without a standardized protocol.12,13,15,19 A recent prospective study reported an FNR of 37%.20 However, the study protocol used only 8 US-guided 14-gauge needle core-cut biopsies, yielding significantly less tissue for analysis compared with the present pooled analysis.
Refinement of the biopsy technique and more careful patient selection is associated with lower FNR. A trial from the MD Anderson Cancer Center showed that in ERBB2-positive and TN breast cancer with a partial or near-complete radiologic response to NACT, only 2 of 40 patients had a false-negative event using image-guided VAB.14 An FNR of 10% was reported from a trial from Seoul National University Hospital, using image-guided biopsy with at least 5 samples in patients with excellent response to treatment as defined by residual tumor or tumor bed measuring 0.5 cm or smaller on post-NACT breast MRI scan.16 A recent prospective study from The Royal Marsden Hospital showed similar results with an FNR of 9.1% for TN and ERBB2-positive disease after post-NACT image-guided VAB. This study included a highly selected group of patients who had sampling of at least 90% of a residual tumor bed measuring 2 cm or smaller.17 Use of VAB with 7-gauge needle was also shown to be associated with an FNR less than 10% in a large multicenter prospective study.19
The exploratory analysis in this study showed that careful patient selection and application of a standardized image-guided biopsy protocol was associated with improved diagnostic accuracy. The selected subgroup included patients who at completion of NACT had a residual imaging tumor bed measuring 2 cm or smaller and underwent image-guided VAB with at least 6 representative samples. Application of this diagnostic algorithm across all subtypes resulted in an FNR of 3.2% and an NPV of 97.4%. Specifically, in the subgroup of patients with ERBB2-positive and TN breast cancer, which are the most likely to achieve pCR, post-NACT VAB showed an FNR of 4.2% and an NPV of 97.2% (Table 4). Because the accuracy of the technique is dependent on the ability to reliably sample the area of residual imaging abnormality/tumor bed, the use of VAB was selected for the subgroup analysis because it allows retrieval of larger tissue samples compared with core-cut biopsy.18 A single 10-gauge needle VAB retrieves approximately 221 mg of tissue. This is the equivalent of thirteen 14-gauge needle core-cut biopsies. In the MD Anderson Cancer Center feasibility trial on the use of image-guided biopsy to identify exceptional responders, the 2 patients with false-negative events had 4 and 6 samples taken, respectively (the median number of samples for the study was 12).14 Results from the Seoul National University Hospital trial support an image-guided biopsy protocol with 5 or more samples.16 Based on these findings, a minimum number of 6 biopsies was set for the exploratory analysis of this study because this was also the median number of biopsy cores in the present cohort. Finally, a maximum size of 2 cm for the residual imaging abnormality was selected because this represents the cutoff for T1 tumors based on TNM staging system and could be easily included in diagnostic algorithms and used by the wider medical community. Moreover, a size of 2 cm was considered by participating radiologists as optimal to allow adequate sampling without narrowing patient selection. A similar diagnostic protocol, with use of a minimum of twelve 9-gauge image-guided VABs of a residual tumor bed of 2 cm or smaller, is currently used in a prospective clinical trial at the MD Anderson Cancer Center, assessing the safety of elimination of surgery in exceptional responders (NCT02945579).
Surgery after NACT is still considered standard of care. However, this is often not guided by response to treatment.21 With the use of modern chemotherapy, anti-ERBB2 therapy regimens, and recent immunotherapy, high pCR rates exceeding 60% have been reported, especially in TN and ERBB2-positive breast cancer.4,5,6,22 This suggests that the role of surgery is limited to histopathologic confirmation of pCR for a significant proportion of patients. In this setting, surgical intervention might seem counterintuitive to efforts for tailored treatments because it exposes patients to potentially unnecessary procedures with their associated morbidity and psychological burden. However, the challenge is to reliably identify those patients who have achieved pCR and may be spared surgery.
The results of this multicenter pooled data analysis emphasize the importance of using a standardized image-guided biopsy protocol as a tool to assess the presence of residual cancer in the breast at the end of NACT and prior to any surgical intervention. With an FNR of less than 5%, this technique may reliably predict pCR and identify individuals who may not benefit from further local treatment. Moreover, it was observed that in patients with breast pCR, the likelihood of nodal disease ranged between 3.6% and 10.3% in those presenting with clinically node-negative and node-positive disease, respectively. This is further supported by published data suggesting that pCR in the breast is highly correlated with the nodal status after NACT especially in patients with TN and ERBB2-positive subtypes.8,23 The risk of residual nodal disease is so low, particularly in the subgroup of patients who were clinically node-negative at diagnosis, that consideration of omission of breast and axillary surgery in the context of clinical trials, in a selected group of patients who have no residual disease on post-NACT image-guided breast biopsy and axillary imaging complete response, warrants further consideration.
Limitations
Although this study was well designed and conducted and represents a multicenter pooled analysis of patient-level data from a large cohort from 3 large comprehensive cancer centers, there are potential limitations. Image-guided biopsy to assess the presence of residual disease in the breast after NACT and before surgery is not routinely performed and not included in standard breast cancer management pathways. The technique is operator dependent and requires breast radiologist expertise using standardized assessment protocols. Therefore, the findings of this pooled analysis can only be carefully implemented in specialized centers and at this time only in the context of clinical trials in appropriately defined and selected patients.
Conclusions
The results of this large, multicenter pooled analysis of patient-level data suggest that a standardized protocol using image-guided VAB, to obtain at least 6 representative samples of a residual breast imaging abnormality or tumor bed measuring 2 cm or smaller, can reliably identify patients with no residual disease in the breast following NACT. These findings support the use of this technique to predict response to treatment and should inform the design of future de-escalation trials in exceptional responders.
eAppendix. Inclusion and exclusion criteria for eligibility and data collected
eTable. Histopathology details of patients with no evidence of cancer on image-guided breast biopsy and residual cancer on final surgical histopathology in whole cohort
References
- 1.Killelea BK, Yang VQ, Mougalian S, et al. Neoadjuvant chemotherapy for breast cancer increases the rate of breast conservation: results from the National Cancer Database. J Am Coll Surg. 2015;220(6):1063-1069. doi: 10.1016/j.jamcollsurg.2015.02.011 [DOI] [PubMed] [Google Scholar]
- 2.Masuda N, Lee SJ, Ohtani S, et al. Adjuvant capecitabine for breast cancer after preoperative chemotherapy. N Engl J Med. 2017;376(22):2147-2159. doi: 10.1056/NEJMoa1612645 [DOI] [PubMed] [Google Scholar]
- 3.von Minckwitz G, Huang CS, Mano MS, et al. ; KATHERINE Investigators . Trastuzumab emtansine for residual invasive HER2-positive breast cancer. N Engl J Med. 2019;380(7):617-628. doi: 10.1056/NEJMoa1814017 [DOI] [PubMed] [Google Scholar]
- 4.Sikov WM, Berry DA, Perou CM, et al. Impact of the addition of carboplatin and/or bevacizumab to neoadjuvant once-per-week paclitaxel followed by dose-dense doxorubicin and cyclophosphamide on pathologic complete response rates in stage II to III triple-negative breast cancer: CALGB 40603 (Alliance). J Clin Oncol. 2015;33(1):13-21. doi: 10.1200/JCO.2014.57.0572 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Schneeweiss A, Chia S, Hickish T, et al. Pertuzumab plus trastuzumab in combination with standard neoadjuvant anthracycline-containing and anthracycline-free chemotherapy regimens in patients with HER2-positive early breast cancer: a randomized phase II cardiac safety study (TRYPHAENA). Ann Oncol. 2013;24(9):2278-2284. doi: 10.1093/annonc/mdt182 [DOI] [PubMed] [Google Scholar]
- 6.van Ramshorst MS, van der Voort A, van Werkhoven ED, et al. ; Dutch Breast Cancer Research Group (BOOG) . Neoadjuvant chemotherapy with or without anthracyclines in the presence of dual HER2 blockade for HER2-positive breast cancer (TRAIN-2): a multicentre, open-label, randomised, phase 3 trial. Lancet Oncol. 2018;19(12):1630-1640. doi: 10.1016/S1470-2045(18)30570-9 [DOI] [PubMed] [Google Scholar]
- 7.Boughey JC, Suman VJ, Mittendorf EA, et al. ; Alliance for Clinical Trials in Oncology . Sentinel lymph node surgery after neoadjuvant chemotherapy in patients with node-positive breast cancer: the ACOSOG Z1071 (Alliance) clinical trial. JAMA. 2013;310(14):1455-1461. doi: 10.1001/jama.2013.278932 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Barron AU, Hoskin TL, Day CN, Hwang ES, Kuerer HM, Boughey JC. Association of low nodal positivity rate among patients with ERBB2-positive or triple-negative breast cancer and breast pathologic complete response to neoadjuvant chemotherapy. JAMA Surg. 2018;153(12):1120-1126. doi: 10.1001/jamasurg.2018.2696 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Caudle AS, Yang WT, Krishnamurthy S, et al. Improved axillary evaluation following neoadjuvant therapy for patients with node-positive breast cancer using selective evaluation of clipped nodes: implementation of targeted axillary dissection. J Clin Oncol. 2016;34(10):1072-1078. doi: 10.1200/JCO.2015.64.0094 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Mak KS, Harris JR. Radiotherapy issues after neoadjuvant chemotherapy. J Natl Cancer Inst Monogr. 2015;2015(51):87-89. doi: 10.1093/jncimonographs/lgv003 [DOI] [PubMed] [Google Scholar]
- 11.van la Parra RFD, Kuerer HM. Selective elimination of breast cancer surgery in exceptional responders: historical perspective and current trials. Breast Cancer Res. 2016;18(1):28. doi: 10.1186/s13058-016-0684-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Heil J, Schaefgen B, Sinn P, et al. Can a pathological complete response of breast cancer after neoadjuvant chemotherapy be diagnosed by minimal invasive biopsy? Eur J Cancer. 2016;69:142-150. doi: 10.1016/j.ejca.2016.09.034 [DOI] [PubMed] [Google Scholar]
- 13.Heil J, Kümmel S, Schaefgen B, et al. Diagnosis of pathological complete response to neoadjuvant chemotherapy in breast cancer by minimal invasive biopsy techniques. Br J Cancer. 2015;113(11):1565-1570. doi: 10.1038/bjc.2015.381 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kuerer HM, Rauch GM, Krishnamurthy S, et al. A clinical feasibility trial for identification of exceptional responders in whom breast cancer surgery can be eliminated following neoadjuvant systemic therapy. Ann Surg. 2018;267(5):946-951. doi: 10.1097/SLA.0000000000002313 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Tasoulis MK, Roche N, Rusby JE, et al. Post neoadjuvant chemotherapy vacuum assisted biopsy in breast cancer: can it determine pathologic complete response before surgery? J Clin Oncol. 2018;36(15_suppl):567-567. [Google Scholar]
- 16.Lee HB, Han W, Kim SY, et al. Prediction of pathologic complete response using image-guided biopsy after neoadjuvant chemotherapy in breast cancer patients selected based on MRI findings: a prospective feasibility trial. Breast Cancer Res Treat. 2020;182(1):97-105. doi: 10.1007/s10549-020-05678-3 [DOI] [PubMed] [Google Scholar]
- 17.Teoh V, MAcNeill F, Roche N, et al. Image-guided vacuum-assisted biopsy to assess pathologic complete response in breast cancer patients with exceptional response to neoadjuvant chemotherapy. J Global Oncol. 2019;5(suppl 1):abstr 39. [Google Scholar]
- 18.O’Flynn EA, Wilson AR, Michell MJ. Image-guided breast biopsy: state-of-the-art. Clin Radiol. 2010;65(4):259-270. doi: 10.1016/j.crad.2010.01.008 [DOI] [PubMed] [Google Scholar]
- 19.Heil J, Pfob A, Sinn H-PP, et al. Abstract GS5-03: diagnosing residual disease and pathologic complete response after neoadjuvant chemotherapy in breast cancer patients by image-guided vacuum-assisted breast biopsy: results of a prospective multicenter trial. Cancer Res. 2020;80(4 suppl):GS5-03-GS05-03. [Google Scholar]
- 20.Peeters M-JTFDV, van Loevezijn A, van der Noordaa ME, et al. Abstract GS5-06: towards omitting breast surgery in patients with a pathologic complete response after neoadjuvant systemic treatment: interim analysis of the MICRA trial (Minimally Invasive Complete Response Assessment). Cancer Res. 2020;80(4 suppl):GS5-06-GS05-06. [Google Scholar]
- 21.Karakatsanis A, Tasoulis MK, Wärnberg F, Nilsson G, MacNeill F. Meta-analysis of neoadjuvant therapy and its impact in facilitating breast conservation in operable breast cancer. Br J Surg. 2018;105(5):469-481. doi: 10.1002/bjs.10807 [DOI] [PubMed] [Google Scholar]
- 22.Schmid P, Cortés J, Dent R, et al. LBA8_PRKEYNOTE-522: phase III study of pembrolizumab (pembro) + chemotherapy (chemo) vs placebo (pbo) + chemo as neoadjuvant treatment, followed by pembro vs pbo as adjuvant treatment for early triple-negative breast cancer (TNBC). Ann Oncol. 2019;30(suppl 5). [Google Scholar]
- 23.Tadros AB, Yang WT, Krishnamurthy S, et al. Identification of patients with documented pathologic complete response in the breast after neoadjuvant chemotherapy for omission of axillary surgery. JAMA Surg. 2017;152(7):665-670. doi: 10.1001/jamasurg.2017.0562 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eAppendix. Inclusion and exclusion criteria for eligibility and data collected
eTable. Histopathology details of patients with no evidence of cancer on image-guided breast biopsy and residual cancer on final surgical histopathology in whole cohort
