Clinical treatment of breast cancer is successful. What remains a challenge for individualized therapy is the prediction of which cancer responds to which treatment. Emerging evidence indicates that mechanical alterations on a (sub-) cellular level are markers of cancer aggressiveness. Detecting these could serve as biomarker and help optimize cancer diagnosis, therapy choice, and patient follow-up. We developed a diagnostic platform called ARTIDIS (Automated and Reliable Tissue Diagnostics) based on atomic force microscopy (AFM)-type nanomechanical testing. ARTIDIS uses a 20-nanometer-sharp tip measuring 10'000 miniscule indentation across the biopsy's surface and quantifies stiffness on a molecular level. The relative size and distribution of stiffness values is integrated into a data analytics tool that systematically analyses tissue biomechanics in correlation to clinical parameters, histological prognostic markers, and treatment yielding prognostic and predictive value.

To demonstrate ARTIDIS' clinical utility for diagnosis and development of prognostic and predictive biomarkers we conducted a prospective blinded clinical study on 545 patients. All patients undergoing a core needle or vacuum biopsy of the breast qualified. Within the routine clinical workflow, ARTIDIS analyzed fresh biopsies in a physiological solution preserving the biopsy's viability for histological and genetic analysis. Per patient, one representative biopsy was measured, marked and reunited with the patient's other biopsies for standard histological procedure. In total, 583 biopsies from N=545 patients were analyzed: 62.1% were B1-B4 (normal, benign, of uncertain malignant potential, or suspicious); a third (36.9%) malignant (B5a-B5d), and six specimens (1%) not assignable. Half of all B5b lesions were Luminal B, followed by Luminal A, Luminal B-like, HER2-positive or triple negative. The analysis of the primary endpoint (N=520 patients qualified, N=3 were excluded due to missing information) demonstrated ARTIDIS' ability to detect cancer in all samples including lesions with <5% neoplastic tissue: diagnostic ROC curve of B1+B2 lesions vs. all B5 lesions (CI 95%; 96% sensitivity, 78% specificity, AUC = 0.94) . Secondary endpoints were evaluated and included ARTIDIS' ability to distinguish and further categorize molecular subtypes of breast cancer.

We could separate more aggressive cases within the Luminal B subtype (CI 95%; sensitivity 83%, specificity 82%, AUC = 0.86). Secondary analysis also indicated a similarity between the majority of B3 (uncertain) and B2 (benign) lesions and distinct differences from all B5a and B5b, which could help reduce overtreatment of B3 lesions.

This large prospective single center study demonstrated ARTIDIS' ability to differentiate benign from malignant lesions and to further subclassify specific breast cancer subtypes. In addition, Luminal B cancers exhibited nanomechanical profiles associated with better or worse prognosis and treatment outcome. The final analysis results provide cutoffs that will be assessed in the upcoming large multicenter prospective trial. Finally, this study helped to determine clinical utility of nanomechanical characterization of breast and other solid cancers for prediction and treatment optimization.

Citation Format: Rosemarie Burian, Ahmed Jizawi, Flora Scott, Sabine Schädelin, Christian Raez, Simone Muenst-Soysal, Tatjana Vlaijnic, Ellen Obermann, Sophie Dellas, Serafino Forte, Zlatko Marušić, Tobias Appenzeller, Philipp Oertle, Nam Seung-Zln, Gregory Zaugg, Milan Liepelt, Gabriel Zihlman, Roderick Lim, Marko Loparic, Marija Plodinec. First clinical validation of the physical biomarker based on the nanomechanical tissue profiling for rapid breast cancer diagnosis, prognosis and predication of treatment outcome [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr LB-273.