Click here to

Session: Therapy: Outcome Modeling and Assessment I [Return to Session]

Prediction of Therapeutic Outcomes of Patients with Non-Small Cell Lung Cancer for Stereotactic Ablative Radiotherapy Based On a Differential Model

E Mai*1, A Hidetada2,T Huy1,Y Shirakawa3, N Kenta1, Y Tadamasa4,S Yoshiyuki5,(1) Division of Medical Quantum Science, Department of Health Sciences, Graduate School of Medical Science, Kyushu University, Fukuoka,JP(2)Division of Medical Quantum Science, Department of Health Sciences, Faculty of Medical Sciences, Kyushu University,Fukuoka,JP(3)Department of Radiation Oncology, National Hospital Organization Kyushu Cancer Center,Fukuoka,JP(4)Department of Radiology, Kyushu University Hospital,Fukuoka,JP(5)Kyushu International Heavy Ion Cancer Treatment Center,Saga,JP


MO-IePD-TRACK 6-6 (Monday, 7/26/2021) 12:30 PM - 1:00 PM [Eastern Time (GMT-4)]

Purpose: For unresectable non-small cell lung cancer (NSCLC) patients at a Stage 1 or 2, stereotactic ablative radiotherapy (SART) is the first treatment choice alongside surgery. A five-year survival rate for patients received SART is reportedly about 70%, whose patients may have been suitable for SART. Therefore, the evaluation approaches to choose suitable patients for SART could be needed. We aimed to predict radiotherapeutic outcomes of patients with NSCLC from computed tomography (CT) scans using a differential model.

Methods: Five NSCLC patients receiving SART that had planning CT and follow-up CT images [total scans: 23 (4.6 scan/patient)] were employed for extraction of GTVs and calculation of tumor cell numbers. The number of tumor cells at a time (CT scan) for each tumor, except CT images with radiation pneumonitis, was calculated by multiplying the density of tumor cells with the tumor volume obtained from the CT images. The tumor growth trajectories with the time-variant number of tumor cells were predicted by a Gompertz model with a linear-quadratic model under the assumption that each tumor cell could be resistant or sensitive to radiation. Three parameters in the models (degrees of radiation resistance and sensitivity, and the ratio of two types of cells) were optimized by a Levenberg-Marquardt method. The proposed method was evaluated with the mean absolute percentage error (MAPE) between predicted and reference numbers of tumor cells.

Results: The MAPE ranged from 6.09 to 83.10%, and the average MAPE was 29.3%.

Conclusion: This study showed the potential of predicting the therapeutic effect after SART.



    Modeling, Radiation Therapy, Linear Quadratic Model


    Not Applicable / None Entered.

    Contact Email