2. UroToday Home
  3. Conference Highlights
  4. ASCO 2024
  5. ASCO 2024 Prostate Cancer

ASCO 2024

ASCO 2024: Prognostic Validation of a Digital Pathology-Based Multi-Modal Artificial Intelligence Biomarker in Patients with Metastatic Hormone-Sensitive Prostate Cancer from the CHAARTED Trial (ECOG-ACRIN EA3805)

(UroToday.com) The 2024 American Society of Clinical Oncology (ASCO) annual meeting held in Chicago, IL between May 31 and June 4 was host to the Poster Session: Genitourinary Cancer: Prostate, Testicular, and Penile. Dr. Mark Markowski presented the results of the prognostic validation of a digital pathology-based multi-modal artificial intelligence (MMAI) biomarker in patients with metastatic hormone-sensitive prostate cancer (mHSPC) from the CHAARTED trial (ECOG-ACRIN EA3805)


The CHAARTED trial (ChemoHormonal Therapy versus Androgen Ablation Randomized Trial for Extensive Disease in Prostate Cancer) investigated the efficacy of combining docetaxel chemotherapy with androgen-deprivation therapy (ADT) compared to ADT alone in patients with metastatic hormone-sensitive prostate cancer (mHSPC).1

The ArteraAI digital pathology-based multi-modal artificial intelligence (MMAI) prognostic biomarker utilizes biopsy digital histopathology images and clinical data to stratify the risk of patients with localized prostate cancer and prognosticate the risk of distant metastasis. Recently, ArteraAI Prostate has been developed and validated in localized prostate cancer to provide more precise risk stratification for multiple endpoints compared to the NCCN risk group classification.2

The objective of this study was to validate the ArteraAI MMAI biomarker in mHSPC. This validation was conducted using a subset of patients from the CHAARTED trial (n=456/790, 57.7%), aiming to assess the biomarker's effectiveness in predicting outcomes in this specific patient population. Patients included were required to have the following data: age at diagnosis, baseline PSA, T-stage, and available digital histopathology images either from biopsy or prostatectomy to validate the MMAI. 

The patient's baseline characteristics were compared to examine the balance between included and excluded patients: Patients with and without digital pathology available (DPEP vs non-DPEP). The characteristics are illustrated in the table below. Additionally, using the NCCN guidelines,3 the investigators defined four prognostic groups based on combinations of tumor volume (high (HV) vs. low (LV)) and synchronous (S) vs metachronous (M) disease.

  1. LV-M
  2. HV-M
  3. LV-S
  4. HV-S


Using the existing MMAI model, scores were generated, and its prognostic ability was evaluated for overall survival (OS) in Cox proportional hazard models. These clinical groups were evaluated using multivariable analysis (MVA) which included the MMAI risk group and treatment.

The median follow-up of the censored patients was 4.1 (IQR: 3.3-5.0) years and the estimated 5-yr OS across MMAI high, intermediate, and low groups was 39%, 58%, and 83%, respectively (log-rank p=<0.001). MMAI score was prognostic for OS (HR: 1.5 (95% Cl: 1.33-1.73) p<0.001; per SD increase) on univariable analysis.image-1.jpg
Of the included patients, 370 (81.1%) were classified as MMAI-high and 86 (18.9%) as MMAI-intermediate/low risk. Data were available from 394/456 patients for classification into the four subgroups mentioned above:

  • LV-M (N=57) MMAI-high 56.1%
  • HV-M (N=29) MMAI-high 69%
  • LV-S (N=66) MMAI-high 86.4%
  • HV-S (N=242) MMAI-high 92.6%

On multivariate analysis, the MMAI-high-risk group model was prognostic for OS (HR: 1.77 (95% CI: 1.10-2.84) p=0.02) adjusting for treatment arm, volume status (HV vs. LV), and stage at diagnosis (S vs M) as outlined in the table below:
multivariate analysis, the MMAI-high risk group model
Dr. Markowski wrapped up his presentation with the following take-home messages:

  • The ArteraAI MMAI model was found to be prognostic for OS among a subset of men with mHSPC in CHAARTED.
  • The prognostic effect for OS persisted when controlling for treatment, metastatic burden (HV vs. LV), and metastatic status (S vs. M) at diagnosis.
  • Future research will include the development of an MMAl model optimized for advanced prostate cancer and possibly predictive for docetaxel benefit

Presented by: Mark Christopher Markowski, MD, PhD, Genitourinary Medical Oncologist, Johns Hopkins University, Baltimore, MD

Written by: Julian Chavarriaga, MD – Society of Urologic Oncology (SUO) Clinical Fellow at The University of Toronto, @chavarriagaj on Twitter during the 2024 American Society of Clinical Oncology (ASCO) annual meeting held in Chicago, IL between May 31st and June 4th. 

Related content: Prognostic Power of Multi-Modal Artificial Intelligence Biomarker in CHAARTED Trial Subset - Mark Markowski

References:

  1. Kyriakopoulos CE, Chen YH, Carducci MA, Liu G, Jarrard DF, Hahn NM, Shevrin DH, Dreicer R, Hussain M, Eisenberger M, Kohli M, Plimack ER, Vogelzang NJ, Picus J, Cooney MM, Garcia JA, DiPaola RS, Sweeney CJ. Chemohormonal Therapy in Metastatic Hormone-Sensitive Prostate Cancer: Long-Term Survival Analysis of the Randomized Phase III E3805 CHAARTED Trial. J Clin Oncol. 2018 Apr 10;36(11):1080-1087. doi: 10.1200/JCO.2017.75.3657. Epub 2018 Jan 31. PMID: 29384722; PMCID: PMC5891129.
  2. Esteva A, Feng J, van der Wal D, et al. Prostate cancer therapy personalization via multi-modal deep learning on randomized phase III clinical trials. NPJ Digit Med. 2022 Jun 8;5(1):71.
  3. NCCN. NCCN Clinical Practice Guidelines in Oncology Prostate Cancer. 2024; v3.2024.https://www.nccn.org/professionals/physician_gls/pdf/prostate.pdf