Margin status after breast-conserving surgery (BCS) is a critical prognostic factor in breast cancer. Adjunct, intraoperative analysis of excised tumor margins reduces the risk of reexcision surgery by allowing the surgeon to take additional tissue if an involved margin is detected. While intraoperative specimen radiography (SXR) is commonly used as such an adjunct, wide-field optical coherence tomography (WF-OCT) imaging has also demonstrated performance for this purpose. WF-OCT uses near-infrared interferometry to produce high-resolution images of tissue microarchitecture at a depth up to 2 mm and may detect residual malignant tissue not apparent on SXR.
This retrospective case series compared SXR with WF-OCT images, collected, and analyzed prospectively. Adult women undergoing BCS for biopsy-proven invasive ductal carcinoma (IDC) or ductal carcinoma in situ (DCIS) at 2 centers were included. Primary excised tumor specimens were imaged intraoperatively with SXR and WFOCT (Perimeter Medical Imaging AI). Additional shaves were taken based on SXR results. WF-OCT images were blinded and reviewed, and not used for clinical decision-making. After closing, all tissue was sent for permanent histopathology, which was designated as ground truth. Histopathology-positive margins in patients without additional cavity shaves were designated as SXR false negatives (SXR-FNs). The WF-OCT results and images from FN patients were compared with the corresponding SXR and final histopathology images.
Eighty-nine consecutive patients undergoing BCS were imaged with SXR and WF-OCT prior to permanent histopathology. Average sensitivity and specificity of SXR were 76.5% and 69.6%, respectively. Six cases, comprising
4 DCIS, 1 IDC, and 1 DCIS with IDC, were designated SXR-FN (additional shaves not taken intraoperatively; positive margin present on final histopathology). In the 5 SXR-FN DCIS cases, suspicious tissue changes were present on WF-OCT; the 1 SXR-FN case not detected by WF-OCT was IDC. Side-by-side comparison of SXR-FN, WFOCT, and ground-truth histology shows the correlating lesion features.
This case series demonstrates that, consistent with ground-truth permanent histology, WF-OCT is able to detect margin positivity in excised tissue that was not apparent on SXR, particularly with respect to DCIS. Although a small case series, these results are encouraging. Additional studies should be tested in a randomized controlled trial.
Savitri Krishnamurthy,1 Payal Salgia,2 David Rempel,2 Andrew Berkeley,2 Beryl Augustine,2 Chandandeep Nagi,3 Alastair Thompson3
1The University of Texas MD Anderson Cancer Center, Houston, TX
2Perimeter Medical Imaging AI, Toronto, Ontario, Canada
3Baylor College of Medicine, Houston, TX