Digital Breast Tomosynthesis (DBT) to Improve Assessment of Resection Margins in Invasive Breast Cancer

Overview

The challenge of breast conserving surgery (BCS) is to obtain free resection margins, by removing the invasive tumor and the precursor cells: DCIS (Ductal Carcinoma in Situ). Consequently, 17%-25% of primary invasive breast cancer patients will need a repeat surgery. Repeat breast surgery has been associated with higher surgical risks, poorer cosmetic outcome and increased psychological and economic burden. Finding a precise method to obtain tumor- and DCIS-free resection margins during BCS is therefore strongly needed. The purpose with this study is to investigate if the need for a re-operation can be reduced by performing Digital Breast Tomosynthesis (DBT) of the resected tumor margins compared to the currently used standard X-ray during BCS.

Background Breast cancer is the most common type of cancer in Denmark with approximately 4700 new cases each year, with an increased incidence during the last decades. Almost 70% of the patients are treated with breast conserving surgery (BCS) and breast irradiation. Unfortunately, 5.9 % of the patients treated with BCS will experience local recurrence after 9 years. The strongest predictor for local recurrence is a positive pathologic resection margin. Therefore it is important intraoperatively to be able to evaluate the margin width of the resected specimen precisely. The challenge during BCS is to remove the primary invasive breast cancer, with negative resection margin, free of invasive tumor and DCIS, Ductal Carcinoma In Situ. Some have concluded that lack of accurate imaging especially of the DCIS is the main reason for positive margins and repeat surgery, as the DCIS often, but not in all cases, contains microcalcifications that can be detected on mammography (x-ray of the breast). Consequently, 17%-25% of the patients will need a second or a third surgery. Repeat surgery has been associated with higher surgical risk, poorer cosmetic outcome, and increased psychological and economic burden for both patients and the society. Finding an optimal imaging method to obtain adequate resection margins close to or during BCS that can reduce the number of repeat surgeries, is therefore strongly needed. Definition of resection margin status A safe macroscopic surgical resection margin is, according to the Danish Breast Cancer Group´s (DBCG) guidelines, defined as a margin width ≥ 5mm minimizing the risk of leaving residual tumor. In many institutions in Denmark, a perioperative evaluation of the tumor is performed as a macroscopically assessment by a pathologist. This is a ressource demanding procedure and X-ray of the specimen is used instead. A radiological, positive resection margin is defined as a margin width of 0 mm, that leads to a re-excision during the same surgical procedure, and additional breast tissue is removed at the exact same area of the patient's breast. This procedure aims to avoid the discomfort of a repeat surgery. The perioperative radiological assessment of the resection margins is afterwards confirmed microscopically, where a positive margin at our institution is defined as "tumor on ink", meaning tumor cells in the inked resection margin with 0 mm to the margin. A positive margin will trigger a repeat surgery, either a new BCS or a mastectomy where the whole breast is removed. For DCIS, a margin-width of less than 2 mm is considered positive, and the patient will be offered repeat surgery according to guidelines. A distance from the invasive carcinoma to the resection margin of \< 2 mm, but not "on the ink" is surgical considered sufficient and the patient will according to DBCG guidelines, not require a repeat surgery, but will instead be offered additional boost to the subsequent radiotherapy. The result of the final histopathology is, however, not available until 7 workdays after the initial BCS. Guidance of a more optimal perioperative methods to evaluate the resection margins during BCS would, therefore, be highly desirable, and could potentially lead to a decreased rate of repeat surgery and hence save several patients from the discomfort of a second surgical procedure. Intraoperative imaging methods for margin assessment Various methods to assess resection margins have been investigated such as frozen section analysis, imprint cytology, macroscopically evaluation and imaging with x-ray, micro-CT scanning or ultrasound of the resected tumor margins. All have, however, their limitations, either because they are time consuming, or simply because they are not accurate enough to predict resection margin status. The imaging modality X-ray has been reported in several studies to have a high specificity, predicting negative margins up to 95%, but unfortunately a low sensitivity of only 21%. That means that the method is particularly good to predict a negative margin. The X-ray is less sensitive when it comes to predict a positive resection margin, with the final histopathology as a reference method. The lower sensitivity is due to the fact, that x-ray has difficulties, when it comes to detect the breast tumor in a very dense normal breast tissue, most common for young women. The X-ray evaluation also has the disadvantage of only delivering a 2D solution to a 3D problem. The other imaging method currently in use is ultrasound that has a high sensitivity of 80% but a low specificity of 59% to predict positive resection margins (Ramos M. et al). A new method called Micro-CT (micro-Computed Tomography) can provide 3D imaging of the resected tumor and the microcalcifications that helps to asses all margins of the resected tumor. Unfortunately, there is a high interobserver variability reported with this method. Furthermore, it is difficult to differentiate between tumor and normal breast tissue in the resected margins in some of the cases, which makes it impossible for the radiologist to determine if the margins are positive or negative. This may be one of the reasons for, why this method does not perform so well with a sensitivity range of only 38%-56%. Digital Breast Tomosynthesis (DBT) is a novel method for assessment of resection margins based on X-ray technology. DBT provides images of the resected tumor with a large field of view of the specimen and a high resolution, that allows the radiologist to examine the resected specimen in one millimeter increments both in 2D and 3D, where the extent and the shape of the tumor and the microcalcifications are better defined than on conventional 2D mammography. Only few clinical studies have investigated DBT for evaluation of resection margins with the most recent study with 98 patients included (Park KU. et al). This study reported a high performance of the DBT with a sensitivity to predict positive margins of 93%, and a specificity of 78%, to predict negative margins, with the final histopathology as a reference method. Objectives The overall objective of the present study is to investigate if adequate resection margins can be obtained with use of DBT of the resected margins during BCS to reduce the rate of repeat surgery and avoid further unnecessary complications and psychological stress for the patients. Hypothesis Does intraoperative DBT improve assessment of the resected margins compared to standard method with X-ray? The final histopathology will be used as a reference method. 1. DBT can contribute to obtain negative resection margins and thereby reduce the need of repeat surgery as compared to X-ray. 2. The performance of DBT predicting resection margin status is more accurate compared to the standard method with X-ray. Study participants The study participants will be recruited at the Department of Plastic and Breast Surgery at Aarhus University Hospital after they have been diagnosed with operable invasive breast cancer and scheduled for BCS. The study participants will receive separate information leaflet about this project that respects the applicable regulatory requirements. The page describing the rights of the patient participating in a health research project will also be handed out to the patient by the breast surgeon. The breast surgeon will obtain the informed consent from the patient and mention in the patient's medical record, that the patient has given consent to participate in this project. Randomization The study participant will be randomized to either the DBT or the standard method (X-ray) by the breast surgeon in the system REDCap®. REDCap® is a secure web platform that can build and manage online databases and surveys that will also be used for the data management of this project. The randomization will be performed by the breast surgeon during the first consultation with the patient, as soon as the patient has given oral and written consent and is included in each trial. Handling of the resected breast tumor The study encompasses tissue being procured as part of the normal diagnostic procedure. The material will, as all tissue collected for diagnostic purposes, be kept in the existing clinical biobank at the Department of Pathology, at Aarhus University Hospital during as well as after the closure of the study. A research biobank will therefore not be established Safety There are no side effects and no risk of radiation for the investigators in the DBT-trial, as the resected tumor will be located inside the DBT or in the X-ray that is a closed system, ex-vivo, with no radiation exposure to the technical staff or to the study patients. Statistics and analysis A sample size of 123 study participants was calculated as necessary to detect a reduction from 20% in the X-ray-group to 7% in the DBT-group with a statistical power of 80% and a 5% level of significance. The investigators expect a few patients may drop-out for different reasons, and estimated that a total number of 250 patients are enough to be included in the DBT-trial. The statistical analyses and calculation will be made primarily by the investigator Irina Palimaru Manhoobi (IPM) using STATA® where after the work will be reviewed by a statistician to ensure a high quality of the statistical work. 1. The occurrence of positive resections margins, re-excisions and repeat surgery will be compared between groups using the Fisher exact test. 2. The performance of the tested and the standard method: The sensitivity, specificity, Positive Predictive Value (PPV) and the Negative Predictive Value (NPV) compared to the reference method will be calculated with the McNemar Chi-Square test. A two-sided p-value of less than 0.05 will be considered statistically significant. Data management The study participants will be identified from the Central Personal Register system (CPR) with permission obtained the 1st of April 2020 from the Danish Data Protection Agency (DDPA) / "Region Midtjyllands interne fortegnelse over forskningsprojekter" with case number 1-16-02-141-20. The informed consent from all study participants gives the investigator, sponsor and the representants of the sponsor, the direct access to the below described clinical data from the medical record (EPJ) Electronical Patient Journal, that is important to perform this project. All clinical data will be transferred anonymously to the secure web application for data management in "RedCap" that is administered by Aarhus University. The investigator IPM is responsible for the data management. The study is in according to the guidelines of the Helsinki Declaration II, GCP guidelines and the Danish Legislation regarding clinical trials and legislation regarding personal data.