Application of Artificial Intelligence/Machine Vision & Learning for the Development of a Live Single-cell Phenotypic Biomarker Test to Predict Prostate Cancer Tumor Aggressiveness.

To assess the usefulness and applications of machine vision (MV) and machine learning (ML) techniques that have been used to develop a single cell-based phenotypic (live and fixed biomarkers) platform that correlates with tumor biological aggressiveness and risk stratification, 100 fresh prostate samples were acquired, and areas of prostate cancer were determined by post-surgery pathology reports logged by an independent pathologist. The prostate samples were dissociated into single-cell suspensions in the presence of an extracellular matrix formulation. These samples were analyzed via live-cell microscopy. Dynamic and fixed phenotypic biomarkers per cell were quantified using objective MV software and ML algorithms. The predictive nature of the ML algorithms was developed in two stages. First, random forest (RF) algorithms were developed using 70% of the samples. The developed algorithms were then tested for their predictive performance using the blinded test dataset that contained 30% of the samples in the second stage. Based on the ROC (receiver operating characteristic) curve analysis, thresholds were set to maximize both sensitivity and specificity. We determined the sensitivity and specificity of the assay by comparing the algorithm-generated predictions with adverse pathologic features in the radical prostatectomy (RP) specimens. Using MV and ML algorithms, the biomarkers predictive of adverse pathology at RP were ranked and a prostate cancer patient risk stratification test was developed that distinguishes patients based on surgical adverse pathology features. The ability to identify and track large numbers of individual cells over the length of the microscopy experimental monitoring cycles, in an automated way, created a large biomarker dataset of primary biomarkers. This biomarker dataset was then interrogated with ML algorithms used to correlate with post-surgical adverse pathology findings. Algorithms were generated that predicted adverse pathology with >0.85 sensitivity and specificity and an AUC (area under the curve) of >0.85. Phenotypic biomarkers provide cellular and molecular details that are informative for predicting post-surgical adverse pathologies when considering tumor biopsy samples. Artificial intelligence ML-based approaches for cancer risk stratification are emerging as important and powerful tools to compliment current measures of risk stratification. These techniques have capabilities to address tumor heterogeneity and the molecular complexity of prostate cancer. Specifically, the phenotypic test is a novel example of leveraging biomarkers and advances in MV and ML for developing a powerful prognostic and risk-stratification tool for prostate cancer patients.

Clinical Proof-of-concept of a Novel Platform Utilizing Biopsy-derived Live Single Cells, Phenotypic Biomarkers, and Machine Learning Toward a Precision Risk Stratification Test for Prostate Cancer Grade Groups 1 and 2 (Gleason 3 + 3 and 3 + 4).

OBJECTIVE: To examine the ability of a novel live primary-cell phenotypic (LPCP) test to predict postsurgical adverse pathology (P-SAP) features and risk stratify patients based on SAP features in a blinded study utilizing radical prostatectomy (RP) surgical specimens. METHODS: Two hundred fifty-one men undergoing RP were enrolled in a prospective, multicenter (10), and proof-of-concept study in the United States. Fresh prostate samples were taken from known areas of cancer in the operating room immediately after RP. Samples were shipped and tested at a central laboratory. Utilizing the LPCP test, a suite of phenotypic biomarkers was analyzed and quantified using objective machine vision software. Biomarkers were objectively ranked via machine learning-derived statistical algorithms (MLDSA) to predict postsurgical adverse pathological features. Sensitivity and specificity were determined by comparing blinded predictions and unblinded RP surgical pathology reports, training MLDSAs on 70% of biopsy cells and testing MLDSAs on the remaining 30% of biopsy cells across the tested patient population. RESULTS: The LPCP test predicted adverse pathologies post-RP with area under the curve (AUC) via receiver operating characteristics analysis of greater than 0.80 and distinguished between Prostate Cancer Grade Groups 1, 2, and 3/Gleason Scores 3 + 3, 3 + 4, and 4 + 3. Further, LPCP derived-biomarker scores predicted Gleason pattern, stage, and adverse pathology with high precision-AUCs>0.80. CONCLUSION: Using MLDSA-derived phenotypic biomarker scores, the LPCP test successfully risk stratified Prostate Cancer Grade Groups 1, 2, and 3 (Gleason 3 + 3 and 7) into distinct subgroups predicted to have surgical adverse pathologies or not with high performance (>0.85 AUC).

Live-cell phenotypic-biomarker microfluidic assay for the risk stratification of cancer patients via machine learning.

The risk stratification of prostate cancer and breast cancer tumours from patients relies on histopathology, selective genomic testing, or on other methods employing fixed formalin tissue samples. However, static biomarker measurements from bulk fixed-tissue samples provide limited accuracy and actionability. Here, we report the development of a live-primary-cell phenotypic-biomarker assay with single-cell resolution, and its validation with prostate cancer and breast cancer tissue samples for the prediction of post-surgical adverse pathology. The assay includes a collagen-I/fibronectin extracellular-matrix formulation, dynamic live-cell biomarkers, a microfluidic device, machine-vision analysis and machine-learning algorithms, and generates predictive scores of adverse pathology at the time of surgery. Predictive scores for the risk stratification of 59 prostate cancer patients and 47 breast cancer patients, with values for area under the curve in receiver-operating-characteristic curves surpassing 80%, support the validation of the assay and its potential clinical applicability for the risk stratification of cancer patients.

Rapid and Short-term Extracellular Matrix-mediated In Vitro Culturing of Tumor and Nontumor Human Primary Prostate Cells From Fresh Radical Prostatectomy Tissue.

OBJECTIVE: To culture prostate cells from fresh biopsy core samples from radical prostatectomy (RP) tissue. Further, given the genetic heterogeneity of prostate cells, the ability to culture single cells from primary prostate tissue may be of importance toward enabling single-cell characterization of primary prostate tissue via molecular and cellular phenotypic biomarkers. METHODS: A total of 260 consecutive tissue samples from RPs were collected between October 2014 and January 2016, transported at 4°C in serum-free media to an off-site central laboratory, dissociated, and cultured. A culture protocol, including a proprietary extracellular matrix formulation (ECMf), was developed that supports rapid and short-term single-cell culture of primary human prostate cells derived from fresh RP samples. RESULTS: A total of 251 samples, derived from RP samples, yielded primary human tumor and nontumor prostate cells. Cultured cells on ECMf exhibit (1) survival after transport from the operating room to the off-site centralized laboratory, (2) robust (>80%) adhesion and survival, and (3) expression of different cell-type-specific markers. Cells derived from samples of increasing Gleason score exhibited a greater number of focal adhesions and more focal adhesion activation as measured by phospho-focal adhesion kinase (Y397) immunofluorescence when patient-derived cells were cultured on ECMf. Increased Ki67 immunofluorescence levels were observed in cells derived from cancerous RP tissue when compared to noncancerous RP tissue. CONCLUSION: By utilizing a unique and defined extracellular matrix protein formulation, tumor and nontumor cells derived from primary human prostate tissue can be rapidly cultured and analyzed within 72 hours after harvesting from RP tissue.

α-Actinin links extracellular matrix rigidity-sensing contractile units with periodic cell-edge retractions.

During spreading and migration, the leading edges of cells undergo periodic protrusion-retraction cycles. The functional purpose of these cycles is unclear. Here, using submicrometer polydimethylsiloxane pillars as substrates for cell spreading, we show that periodic edge retractions coincide with peak forces produced by local contractile units (CUs) that assemble and disassemble along the cell edge to test matrix rigidity. We find that, whereas actin rearward flow produces a relatively constant force inward, the peak of local contractile forces by CUs scales with rigidity. The cytoskeletal protein α-actinin is shared between these two force-producing systems. It initially localizes to the CUs and subsequently moves inward with the actin flow. Knockdown of α-actinin causes aberrant rigidity sensing, loss of CUs, loss of protrusion-retraction cycles, and, surprisingly, enables the cells to proliferate on soft matrices. We present a model based on these results in which local CUs drive rigidity sensing and adhesion formation.