Introduction

Patient-derived tumor organoids (PDOs) have emerged as a powerful in vitro tool that can estimate anticancer drug response individual patients. EML4?ALK fusion tyrosine kinase was found in 4 to 5% of non?small-cell lung cancers(NSLC), but there was few research using ALK-positive lung cancer organoids. This study aimed to explore a more rapid drug response prediction model, using tumor-organoids in NSLC with presence of ALK rearrangements.

Methods

Tumor tissues were obtained from eight patients who had been diagnosed with NSLC at Seoul ST. Mary?s Hospital. The cancer cell mixture was distributed onto a 384-plate, and organoids culture were done for 3 days, followed by ALK targeted drug exposure. The drug sensitivity prediction of the CODRP (Cancer Organoid based Diagnosis Reactivity Prediction) assay was evaluated by analyzing the response to 3 types of ALK-targeted drugs and comparing them with patient?s clinical records.

Results

Among the 8 patients, there were 5 ALK-positive tissues and 3 ALK-negative tissues. Drug response evaluation for PDOs was performed by Z-score method using quantified AUC (area under curve) values of the dose-response curve. Drug resistance and response groups were further identified. We created CODRP model by adding cell growth rate variables with AUC. The results of conventional drug response predicting model using the AUC-Z-score based method didn?t match the patient's response in real clinical practice. However, drug response results derived by CODRP were consistent with the actual clinical response.

Conclusions

We identified that CODRP prediction model using PDOs improves the accuracy of drug response prediction.