> Details

Pfob, André; Sidey-Gibbons, Chris; Rauch, Geraldine; Thomas, Bettina; Schaefgen, Benedikt; Kuemmel, Sherko; Reimer, Toralf; Hahn, Markus; Thill, Marc; Blohmer, Jens-Uwe; Hackmann, John; Malter, Wolfram; Bekes, Inga; Friedrichs, Kay; Wojcinski, Sebastian; Joos, Sylvie; Paepke, Stefan; Degenhardt, Tom; Rom, Joachim; Rody, Achim; Große, Regina; van Mackelenbergh, Marion; Reinisch, Mattea; Karsten, Maria; [...]

Abstract PD7-02: Intelligent vacuum-assisted breast biopsy to identify breast cancer patients with pathologic complete response after neoadjuvant systemic treatment for omission of breast and axillary surgery

Reference
management

Bookmarks

Remove from
bookmarks

Share this on Whatsapp

Close

Bookmarks



You can manage bookmarks using lists, please log in to your user account for this.
  • Media type: E-Article
  • Title: Abstract PD7-02: Intelligent vacuum-assisted breast biopsy to identify breast cancer patients with pathologic complete response after neoadjuvant systemic treatment for omission of breast and axillary surgery
  • Contributor: Pfob, André; Sidey-Gibbons, Chris; Rauch, Geraldine; Thomas, Bettina; Schaefgen, Benedikt; Kuemmel, Sherko; Reimer, Toralf; Hahn, Markus; Thill, Marc; Blohmer, Jens-Uwe; Hackmann, John; Malter, Wolfram; Bekes, Inga; Friedrichs, Kay; Wojcinski, Sebastian; Joos, Sylvie; Paepke, Stefan; Degenhardt, Tom; Rom, Joachim; Rody, Achim; Große, Regina; van Mackelenbergh, Marion; Reinisch, Mattea; Karsten, Maria; [...]
  • imprint: American Association for Cancer Research (AACR), 2022
  • Published in: Cancer Research
  • Language: English
  • DOI: 10.1158/1538-7445.sabcs21-pd7-02
  • ISSN: 1538-7445; 0008-5472
  • Keywords: Cancer Research ; Oncology
  • Origination:
  • Footnote:
  • Description: <jats:title>Abstract</jats:title> <jats:p>Background: Neoadjuvant systemic treatment (NST) elicits a pathologic complete response (pCR, ypT0, ypN0) in 40-70% of women with HER2 positive, triple-negative, and high-proliferative Luminal B breast cancer. These patients may not need surgery as all local tumor has already been eradicated by NST. However, their safe identification prior to surgery is a major challenge: imaging after NST, minimally-invasive biopsies, or combinations of both using narrow patient selection criteria are not accurate enough either because they showed high rates of missed cancer or high rates of missed pCR. Recently, the concept of an intelligent, minimally-invasive, vacuum-assisted biopsy (intelligent VAB) was introduced to identify exceptional responders to NST. The intelligent VAB is a multivariate risk algorithm that uses artificial intelligence techniques to analyze conventional VAB results alongside contextualizing patient, imaging, and tumor information. It showed great potential to reliably identify patients with a pCR in the breast (ypT0). However, the absent integration of the axillary status impairs clinical applicability. In this study, we evaluated the feasibility of an intelligent VAB to identify exceptional responders to NST in the breast and axilla. Methods: We trained, tested, and validated a machine learning algorithm (Extreme Gradient Boosting Tree) using patient, imaging, tumor, and conventional VAB variables to detect residual cancer after NST (ypT+/is or ypN+) prior to surgery. We used data from 318 women with cT1-3, cN0/+, HER2 positive, triple-negative breast or high-proliferative Luminal B breast cancer who underwent VAB before surgery (NCT02948764). We used 10-fold cross-validation to train and test the algorithm which was externally validated using data of an independent, similar trial (NCT02575612). Findings were compared to the histopathologic evaluation of the surgical specimen. False-negative rate (FNR), specificity, and area under the ROC curve (AUROC) were the main outcome measures. Results: In the development set (n=318), mean patient age was 52.5 years and 45.3% (144 of 318) achieved a pCR (ypT0 and ypN0). Using resampling methods, the intelligent VAB showed an FNR of 5.2% (9 of 174, 95% CI 2.4-9.5), a specificity of 37.5% (54 of 144, 95% CI 29.6-45.9), and an AUROC of 0.92 (95% CI 0.90-0.94) in the development set to detect residual cancer (ypT+/is or ypN+) after NST. In the external validation set (n=45), mean patient age was 48.1 years and 44.4% (20 of 45) achieved a pCR. The intelligent VAB showed an FNR of 0% (0 of 25, 95% CI 0.0-13.7), a specificity of 40.0% (8 of 20, 95% CI 19.1-63.9) and an AUROC of 0.91 (95% CI 0.82-0.97). Spiegelhalter’s Z confirmed a well-calibrated model (z score -0.746, P 0.228). FNR of the intelligent VAB was lower compared to imaging after NST, conventional VAB, or combinations of both using narrow patient selection criteria. Conclusion: An intelligent VAB can reliably exclude residual cancer after NST for women with cT1-3, cN0/+, HER2 positive, triple-negative breast or high-proliferative Luminal B breast cancer. The omission of breast and axillary surgery for these exceptional responders may be evaluated in future trials. Trial registration: NCT02948764 and NCT02575612. Funding: German Research Foundation (DFG)</jats:p> <jats:p>Diagnostic Performance ComparisonFalse-negative rate - % (95% CI); no.Specificity - % (95% CI); no.Negative predictive value - % (95% CI); no.Positive predictive value - % (95% CI); no.AUROC - value (95% CI)Development set (n=318)Imaging after NST24.4% (18.0-13.7); 40 of 16452.2% (43.4-61.0); 69 of 13263.3% (53.5-72.3); 69 of 10966.3% (59.1-73.0); 124 of 187-Conventional VAB32.8% (25.8-40.3); 57 of 174100% (97.5-100); 144 of 14471.6% (64.9-77.8); 144 of 201100% (96.9-100); 117 of 117-Imaging after NST + VAB16.7% (11.4-23.2); 28 of 16832.1% (24.4-40.6); 44 of 13761.1% (48.9-72.4); 44 of 7260.1% (56.1-69.1); 140 of 223-VAB + patient selection9.1% (5.0-14.1) 15 of 17036.3% (28.2-45.0); 49 of 13576.6% (64.3-86.2); 49 of 6464.3% (57.9-70.4); 155 of 241-Intelligent VAB (Extreme Gradient Boosting tree)5.2% (2.4-9.6); 9 of 17437.5% (29.6-45.9); 54 of 14485.7% (74.6-93.3); 54 of 6364.7% (58.5-70.6); 165 of 2550.92 (0.90-0.94)External validation (n=45)Imaging after NST24.0% (9.4-45.1%);6 of 2565.0% (40.8-84.6%);13 of 2068.4% (43.4-87.4%);13 of 1973.1% (52.2-88.4%);19 of 26-Conventional VAB28.0% (12.1-49.4%);7 of 25100% (83.2-100%);20 of 2074.1% (53.7-88.9%);20 of 27100% (81.5-100%);18 of 18-Imaging after NST + VAB12.0% (2.5-31.2); 3 of 2565.0% (40.8-84.6%);13 of 2081.3% (54.4-96.0%); 13 of 1675.9% (56.5-89.7%); 22 of 29-VAB + patient selection4.0% (1.0-2.4); 1 of 2530.0% (9.4-45.1%); 6 of 2085.7% (69.8-99.8); 6 of 763.2% (46.0-78.2); 24 of 38-Intelligent VAB (Extreme Gradient Boosting tree)0.0% (0.0-13.7%);0 of 2540.0% (19.1-63.9%);8 of 20100% (63.1-100%);8 of 867.8% (50.2-82.0%);25 of 370.91 (0.82 - 0.97)AUROC = Area under the receiver operating characteristic curve; CI = confidence interval</jats:p> <jats:p>Citation Format: André Pfob, Chris Sidey-Gibbons, Geraldine Rauch, Bettina Thomas, Benedikt Schaefgen, Sherko Kuemmel, Toralf Reimer, Markus Hahn, Marc Thill, Jens-Uwe Blohmer, John Hackmann, Wolfram Malter, Inga Bekes, Kay Friedrichs, Sebastian Wojcinski, Sylvie Joos, Stefan Paepke, Tom Degenhardt, Joachim Rom, Achim Rody, Regina Große, Marion van Mackelenbergh, Mattea Reinisch, Maria Karsten, Michael Golatta, Joerg Heil. Intelligent vacuum-assisted breast biopsy to identify breast cancer patients with pathologic complete response after neoadjuvant systemic treatment for omission of breast and axillary surgery [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr PD7-02.</jats:p>
  • Access State: Open Access