Intraoperative [18F]FDG flexible autoradiography for tumour margin assessment in breast-conserving surgery: a first-in-human multicentre feasibility study

Abstract

Introduction: In women undergoing breast-conserving surgery (BCS), 20-25% require a re-operation as a result of incomplete tumour resection. An intra-operative technique to assess tumour margins accurately would be a major advantage. A novel method for intraoperative margin assessment was developed by applying a thin flexible scintillating film to specimens-flexible autoradiography (FAR) imaging. A single-arm, multi-centre study was conducted to evaluate the feasibility of intraoperative [¹⁸F]FDG FAR for the assessment of tumour margins in BCS.

Methods: Eighty-eight patients with invasive breast cancer undergoing BCS received ≤ 300 MBq of [¹⁸F]FDG 60-180 min pre-operatively. Following surgical excision, intraoperative FAR imaging was performed using the LightPath^® Imaging System. The first 16 patients were familiarisation patients; the remaining 72 patients were entered into the main study. FAR images were analysed post-operatively by three independent readers. Areas of increased signal intensity were marked, mean normalised radiances and tumour-to-tissue background (TBR) determined, agreement between histopathological margin status and FAR assessed and radiation dose to operating theatre staff measured. Subgroup analyses were performed for various covariates, with thresholds set based on ROC curves.

Results: Data analysis was performed on 66 patients. Intraoperative margin assessment using FAR was completed on 385 margins with 46.2% sensitivity, 81.7% specificity, 8.1% PPV, 97.7% NPV and an overall accuracy of 80.5%, detecting both invasive carcinoma and DCIS. A subgroup analysis based on [¹⁸F]FDG activity present at time of imaging revealed an increased sensitivity (71.4%), PPV (9.3%) and NPV (98.4%) in the high-activity cohort with mean tumour radiance and TBR of 126.7 ± 45.7 photons/s/cm²/sr/MBq and 2.1 ± 0.5, respectively. Staff radiation exposure was low (38.2 ± 38.1 µSv).

Conclusion: [¹⁸F]FDG FAR is a feasible and safe technique for intraoperative tumour margin assessment. Further improvements in diagnostic performance require optimising the method for scintillator positioning and/or the use of targeted radiopharmaceuticals.

Trial registration: Identifier: NCT02666079. Date of registration: 28 January 2016. URL: https://clinicaltrials.gov/ct2/show/NCT02666079 . ISRCTN registry: Reference: ISRCTN17778965. Date of registration: 11 February 2016. URL: http://www.isrctn.com/ISRCTN17778965 .

Keywords: Breast cancer; Breast-conserving surgery; Flexible autoradiography; Margin assessment; Re-operation rate.

Fig. 1

Schematic representation of FAR imaging using a flexible scintillating film. Tumour cells containing a PET-radiotracer (e.g. [¹⁸F]FDG) emit β⁺-particles which are converted to scintillations by the flexible scintillator. As β⁺-particles travel a limited distance in tissue, [¹⁸F]FDG containing cells are detected up to approximately 1 mm in tissue. The scintillations are measured by an ultra-sensitive emCCD camera. The flexible scintillators 12 µm thickness makes the scintillator insensitive to 511 keV ɣ-photons

Fig. 2

Workflow of intraoperative FAR imaging. a The LightPath^® Imaging System was located within the operating theatre. b Immediately following dissection, the intact WLE specimen was placed in a disposable specimen tray. c The specimen tray and specimen were loaded into the light-tight specimen chamber, and a photographic reference image was acquired to confirm that the specimen was correctly positioned. The specimen contours were drawn on a transparent CRF, and each tumour margin was annotated (not shown). d A 5 μm Mylar separator sheet and a flexible scintillator film were draped over the specimen, and a FAR image was acquired. (e) FAR image of WLE specimen shows elevated tumour radiance. The transparent CRF attached to the LightPath^® monitor shows the specimen contours, tumour margin borders, the incision line and the location of the primary tumour (T) in the area of the elevated radiance

Fig. 3

A WLE excision specimen (outlined in yellow) analysed in OsiriX Lite version 11. ROIs are drawn to quantify the signal intensity of the empty background (EBG), tissue background (TBG) and tumour hotspot (THS). The area of increased signal intensity at the superior margin (THS1 superior) contained invasive carcinoma on histopathology analysis

Fig. 4

Receiver operator curve (ROC) of intraoperative FAR performance on intact WLE specimen for various decay corrected injected [¹⁸F]FDG activity thresholds with increments of 1 MBq