Frederic De Leeuw1, Muriel Abbaci1,2, Odile Casiraghi3, Aïcha Ben Lakhdar3, Alexia Alfaro1, Ingrid Breuskin4, Corinne Laplace-Builhé1,2. 1. Gustave Roussy, Plate-forme Imagerie et Cytométrie, Université Paris-Saclay, UMS 23/3655, Villejuif, F-94805, France. 2. UMR CNRS 8081-IR4M, Univ Paris-Sud, Université Paris-Saclay, Orsay, F-91401, France. 3. Département de pathologie, Gustave Roussy, Université Paris-Saclay, Villejuif, F-94805, France. 4. Département de chirurgie, Gustave Roussy, Université Paris-Saclay, Villejuif, F-94805, France.
Abstract
BACKGROUND AND OBJECTIVES: In head and neck surgery, intraoperative and postoperative evaluation of tumor margins is achieved by histopathological assessment, which is a multistep process. Intraoperative analysis of tumor margins to obtain a preliminary diagnosis is usually carried out on frozen sections. Analysis of frozen sections is challenging due to technical difficulties in processing. Full-field optical coherence tomography (FFOCT) provides ex vivo images of fresh tissue samples at a microscopic scale without tissue processing. The objectives of our study were to define the diagnostic criteria required to interpret head and neck FFOCT images and to evaluate the reliability of a histological diagnosis made on an "optical biopsy" produced by head and neck FFOCT imaging compared with conventional histology. STUDY DESIGN/ MATERIALS AND METHODS: First, we established an atlas of comparative images (FFOCT/standard histology) and defined the diagnostic criteria based on FFOCT images. Two pathologists subsequently performed a blinded review on 57 FFOCT images (32 patients). Specificity and sensitivity were measured by comparison with the standard histological diagnosis. The primary endpoint was major concordance, defined as two classifications leading to the same therapeutic decision (treatment/no treatment). RESULTS: Pathologists identified four main criteria for tissue diagnosis on FFOCT images: heterogeneous cell distribution, stromal reaction, coiling, and keratinization abnormalities. The correlation study showed good results, with sensitivity from 88% to 90% and specificity from 81% to 87%, regardless of whether the FFOCT image review was performed by a pathologist with or without previous experience in optical imaging. CONCLUSIONS: Our results demonstrate that FFOCT images can be used by pathologists for differential diagnosis, and that high-resolution FFOCT imaging can provide an assessment of microscopic architecture in head and neck tissues without tissue processing requirements. Lasers Surg. Med.
BACKGROUND AND OBJECTIVES: In head and neck surgery, intraoperative and postoperative evaluation of tumor margins is achieved by histopathological assessment, which is a multistep process. Intraoperative analysis of tumor margins to obtain a preliminary diagnosis is usually carried out on frozen sections. Analysis of frozen sections is challenging due to technical difficulties in processing. Full-field optical coherence tomography (FFOCT) provides ex vivo images of fresh tissue samples at a microscopic scale without tissue processing. The objectives of our study were to define the diagnostic criteria required to interpret head and neck FFOCT images and to evaluate the reliability of a histological diagnosis made on an "optical biopsy" produced by head and neck FFOCT imaging compared with conventional histology. STUDY DESIGN/ MATERIALS AND METHODS: First, we established an atlas of comparative images (FFOCT/standard histology) and defined the diagnostic criteria based on FFOCT images. Two pathologists subsequently performed a blinded review on 57 FFOCT images (32 patients). Specificity and sensitivity were measured by comparison with the standard histological diagnosis. The primary endpoint was major concordance, defined as two classifications leading to the same therapeutic decision (treatment/no treatment). RESULTS: Pathologists identified four main criteria for tissue diagnosis on FFOCT images: heterogeneous cell distribution, stromal reaction, coiling, and keratinization abnormalities. The correlation study showed good results, with sensitivity from 88% to 90% and specificity from 81% to 87%, regardless of whether the FFOCT image review was performed by a pathologist with or without previous experience in optical imaging. CONCLUSIONS: Our results demonstrate that FFOCT images can be used by pathologists for differential diagnosis, and that high-resolution FFOCT imaging can provide an assessment of microscopic architecture in head and neck tissues without tissue processing requirements. Lasers Surg. Med.
Authors: Matti Sievert; Konstantinos Mantsopoulos; Sarina K Mueller; Markus Eckstein; Robin Rupp; Marc Aubreville; Florian Stelzle; Nicolai Oetter; Andreas Maier; Heinrich Iro; Miguel Goncalves Journal: Acta Otorhinolaryngol Ital Date: 2022-02-07 Impact factor: 2.618