Machine Learning Links T-cell Function and Spatial Localization to Neoadjuvant Immunotherapy and Clinical Outcome in Pancreatic Cancer.

Tumor molecular data sets are becoming increasingly complex, making it nearly impossible for humans alone to effectively analyze them. Here, we demonstrate the power of using machine learning (ML) to analyze a single-cell, spatial, and highly multiplexed proteomic data set from human pancreatic cancer and reveal underlying biological mechanisms that may contribute to clinical outcomes. We designed a multiplex immunohistochemistry antibody panel to compare T-cell functionality and spatial localization in resected tumors from treatment-naïve patients with localized pancreatic ductal adenocarcinoma (PDAC) with resected tumors from a second cohort of patients treated with neoadjuvant agonistic CD40 (anti-CD40) monoclonal antibody therapy. In total, nearly 2.5 million cells from 306 tissue regions collected from 29 patients across both cohorts were assayed, and over 1,000 tumor microenvironment (TME) features were quantified. We then trained ML models to accurately predict anti-CD40 treatment status and disease-free survival (DFS) following anti-CD40 therapy based on TME features. Through downstream interpretation of the ML models' predictions, we found anti-CD40 therapy reduced canonical aspects of T-cell exhaustion within the TME, as compared with treatment-naïve TMEs. Using automated clustering approaches, we found improved DFS following anti-CD40 therapy correlated with an increased presence of CD44+CD4+ Th1 cells located specifically within cellular neighborhoods characterized by increased T-cell proliferation, antigen experience, and cytotoxicity in immune aggregates. Overall, our results demonstrate the utility of ML in molecular cancer immunology applications, highlight the impact of anti-CD40 therapy on T cells within the TME, and identify potential candidate biomarkers of DFS for anti-CD40-treated patients with PDAC.

Machine learning links T cell function and spatial localization to neoadjuvant immunotherapy and clinical outcome in pancreatic cancer.

Tumor molecular datasets are becoming increasingly complex, making it nearly impossible for humans alone to effectively analyze them. Here, we demonstrate the power of using machine learning to analyze a single-cell, spatial, and highly multiplexed proteomic dataset from human pancreatic cancer and reveal underlying biological mechanisms that may contribute to clinical outcome. A novel multiplex immunohistochemistry antibody panel was used to audit T cell functionality and spatial localization in resected tumors from treatment-naive patients with localized pancreatic ductal adenocarcinoma (PDAC) compared to a second cohort of patients treated with neoadjuvant agonistic CD40 (αCD40) monoclonal antibody therapy. In total, nearly 2.5 million cells from 306 tissue regions collected from 29 patients across both treatment cohorts were assayed, and more than 1,000 tumor microenvironment (TME) features were quantified. We then trained machine learning models to accurately predict αCD40 treatment status and disease-free survival (DFS) following αCD40 therapy based upon TME features. Through downstream interpretation of the machine learning models' predictions, we found αCD40 therapy to reduce canonical aspects of T cell exhaustion within the TME, as compared to treatment-naive TMEs. Using automated clustering approaches, we found improved DFS following αCD40 therapy to correlate with the increased presence of CD44+ CD4+ Th1 cells located specifically within cellular spatial neighborhoods characterized by increased T cell proliferation, antigen-experience, and cytotoxicity in immune aggregates. Overall, our results demonstrate the utility of machine learning in molecular cancer immunology applications, highlight the impact of αCD40 therapy on T cells within the TME, and identify potential candidate biomarkers of DFS for αCD40-treated patients with PDAC.

Characterization of prostate cancer in needle biopsy by cathepsin B, cell proliferation and DNA ploidy.

BACKGROUND: Our objective was to determine localization patterns of three distinct groups of biomarkers (cathepsin B, MIB-1 and DNA ploidy) in prostate needle biopsy sections to establish localization similarities (or differences) in biopsy and retropubic prostatectomy specimens (RPs). MATERIALS AND METHODS: Prostate needle biopsy specimens and matched RPs from 47 patients with cancer were evaluated. Biopsy and RP sections were stained with anti-cathepsin B (CB) and anti-stefin (cystatin) A (SA) and for cell proliferation and DNA ploidy. The ratio of CB to SA in stained cells was calculated for each biopsy cancer and matched benign prostatic hyperplasia (BPH) sample. RESULTS: The geometric mean of CB to SA was 1.45 in BPH and 2.99 in cancer specimens (p=0.0001). The percentage of S-phase cells and DNA ploidy status in needle biopsy was associated with cancer volume in RP cases (p=0.03). CONCLUSION: Our study has indicated that the ratio of CB to SA is significantly higher in prostate cancer biopsy specimens than in BPH. The percentage of S-phase cells and DNA ploidy in needle biopsies predicts cancer volume of RPs. We have shown that localization of three distinct biomarkers in biopsies reliably assesses the nature of prostate cancer in biopsy sections.

Cathepsin B expression in prostate cancer of native Japanese and Japanese-American patients: an immunohistochemical study.

BACKGROUND: Japanese-American (J-A) men who have immigrated to the U.S.A. and acquired the Western lifestyle usually have more invasive prostate cancer (PCa) than native Japanese (NJ) living in Japan. The specific reasons for these differences remain unknown. The objective of this study was to examine immunostainings of cathepsin B (CB) and its endogenous inhibitor stefin A (SA) in tissue microarray (TMA) and radical prostatectomy (RP) tissue sections in the hope of obtaining insights into the invasiveness of PCa in Japanese patients. PATIENTS AND METHODS: TMA and RP sections were evaluated in 50 men (25 NJ and 25 J-A) for CB and SA reaction products. The CB and SA immunostainings were imaged directly from microscope slides to a computer using a high performance charge coupled device (CCD) digital camera, quantified using Metamorph software, analyzed using the two-sample t-test, and confirmed by multiple regression analysis. RESULTS: The CB and SA proteins were localized in the carcinomatous glands and isolated cancer cells in the TMA and RP sections. The Gleason scores and pre-surgery serum total prostate-specific antigen (PSA) levels did not differ significantly in the NJ and J-A patients (p = 0.14, p = 0.16, respectively). The Chi-square analysis of clinical stage versus place of birth showed that the NJ patients had significantly more T2a and T2b clinical stages than the J-A patients who had more advanced T2c and T3a stages (p = 0.003). The CB and SA immunostainings and their ratios in Gleason score 6 tumors did not show any difference, but the CB:SA ratios in score > or = 7 tumors approached significance levels. CONCLUSION: The overall matching of specimens according to the Gleason grade/score, pre-RP serum total PSA levels, clinical stage and age prior to evaluation of immunostainings greatly minimizes subjectivity associated with the evaluation of markers in this ethnic sub-population of PCa patients. CB and SA immunostaining is similar in Japanese patients who have organ-confined and moderately-differentiated PCa. Analysis of the reaction product data provides indirect evidence that invasiveness of PCa is similar in the two Japanese patient populations.

Heterogeneity of cathepsin B and stefin A expression in Gleason pattern 3+3 (score 6) prostate cancer needle biopsies.

BACKGROUND: There is a significant positive association of increased ratios of cathepsin B to its endogenous inhibitor stefin (cystatin) A in prostatectomy tumors with pelvic lymph node metastases. Needle biopsy diagnosis of prostate cancer is critical in initial treatment selection. The objective was to characterize cathepsin B and stefin A immunostaining patterns in needle biopsies of histologically similar Gleason pattern 3+3 (score 6) foci in relation to pretreatment clinical data. MATERIALS AND METHODS: Immunostaining of cathepsin B and stefin A of 65 biopsy sections were imaged, quantified and analyzed with Student's t-test (p < 0.05). RESULTS: Patients had T1c to T3b clinical stages and pre-surgery total prostate-specific antigen serum levels from 1.25 to 20.0 ng/ml. Cathepsin B and stefin A reaction products were found in the cytoplasm of basal and columnar/cuboidal cells of benign prostatic hyperplasia (BPH), prostatic intraepithelial neoplasia (PIN) and neoplastic cells. Ratios of cathepsin B to stefin A were significantly higher in prostate cancer when compared to that in BPH and PIN glands. CONCLUSION: Small foci of Gleason pattern 3+3 tumors in needle biopsies have heterogeneous cathepsin B and stefin A immunostaining. Stratification of these tumors in relation to clinical stage by cathepsin B and stefin A may assist in treatment selection.