Background and Scope

Our client is a leading CRO based in the US and with customers across US and Europe. The company provides RWD solutions in the form of data subscription and on-demand data collaborations with pharma clients via imaging, genomics, and free text notes. Some of the client’s key RWE offerings include machine learning, model development, biomarker development, clinical trial recruitment, and identification of undiagnosed patients. The client intends to promote value based utilization of advanced real world data sources to support life science companies in developing AI/ML tools for patient care and clinical decision making.

Project Brief

The client had been approached by an AI-based biomarker solution companybased out of South Korea to perform external validation of the NSCLC algorithm to suit different types of cancer. The algorithm was developed to predict whether immunotherapy would be effective for cancer patients. In order to validate the algorithm, the company required a large volume of pathological and clinical data from various cancer patients in immunotherapy (PD-1/PD-L1) in the US and Europe. The company encountered a range of challenges –the most pressing of which was to amass a large volume of advanced real world data. 

Overall Solution

Healthark supported the client in executing the RWE project of sourcing pathological and its associated clinical data of cancer patients on immunotherapy from health system partners in the EU and US.

Patient data was collected and the pathology slides were digitized in the following groups:

Group A

Cancer patients administered with PD-1 and PD-L1 inhibitor indications like squamous cell carcinoma, cervical cancer, Merkel cell carcinoma, Urothelial (bladder) carcinoma, Stomach cancer (including gastric and gastroesophageal junction), Melanoma, Renal cell carcinoma, Hepatocellular carcinoma, Breast Cancer, Endometrial cancer, Small cell lung carcinoma

Group B

Comprised Non-small cell lung cancer (NSCLC) patients were administered PD-1 or PD-L1 inhibitor treatment as 1st line treatment

Group C

Had patients with cancers without PD-1 or PD-L1 inhibitor indications i.e., Pancreatic cancer, Biliary tract cancer, Colorectal cancer, Nasopharyngeal cancer, Ovarian cancer, Prostate cancer, Anal squamous cell carcinoma and other cancer types which were not included in A-B

Healthark’s role

Healthark supported the client in the Biomarker Validation study by providing a range of services:

Overall study design

Conducted literature review to identify key data elements of interest for the study and drafted the feasibility and non-interventional study protocol

SOP development and data extraction

Developed for each site to process clinical and pathological data

 Feasibility assessment assistance

Evaluated the real world data availability from each collaborating site, so as to plan the retrieval of difficult-to-obtain data (e.g. H&E slides of PD-L1 treated patients)

Sites and Principal Investigator (PI) coordination

Identified KOLS and PIs from health system sites, and facilitated the engagement with the PIs to conduct feasibility assessment, and thereby, obtain approval from the Institutional Review Board (IRB) to process data

Project management

Conducted and executed the project end-to-end. This was achieved by creating budgets, project plans and coordinating with the sites to keep track of the project


Healthark supported the client in addressing the RWE needs of an AI company – sourcing “hard to obtain” pathology and clinical data of IO cancer patients and facilitating execution of the entire project. This data aided the validation of an algorithm meant for predicting the likelihood of response therapy of drugs, and thereby, improved the utility of screening tools and treatment decisions.

Share on