Classification of Patients with Painful Tumors to Predict the Response to Palliative Radiotherapy

Abstract

Palliative radiotherapy is an effective and commonly used treatment for pain relief in metastatic cancer patients. This study aimed to classify the patients to predict their response to palliative radiotherapy. This study used the data from our single-center, prospective observational study (UMIN000044984), which included all patients undergoing palliative radiotherapy for the painful tumor, except for postoperative setting or brain metastases. The eligibility criteria were as follows: (1) numerical rating scale (NRS) was two or more before treatment, and (2) underwent palliative radiotherapy between August 2021 and September 2022. Questionnaires for pain site, NRS, and presence/dose of opioids were obtained before, on the last day of treatment, two weeks later, one month later, three months later, and every three months after that. Patients with severe performance status (PS) were contacted by telephone. Pain response was defined to reduce NRS 2 or more at the treated site without an analgesic increase. The primary outcome was the response rate within three months of the treatment. We used multivariable logistic regression to find the independent prognostic factors for the response based on the following covariates: age, sex, PS, body mass index, NRS, history of radiotherapy, treated site, radiographic features (lytic/sclerotic/mixed), metastatic complication, presence/dose of opioids, bone-modifying agents, radiation dose, type of primary tumor, radiosensitivity, number of metastases, and days from diagnosis to treatment. A classification model was created using significant factors, and response rates were calculated for each class. Statistical significance was defined as P<0.05. Among 488 targets of 300 patients registered, 366 targets of 261 patients met the criteria. The median age was 65 years, 44% were female, and PS (0/1/2/3/4) was 6/63/18/9/4%. The major type of tumor was gastrointestinal (32%), lung (19%), and breast (12%). Bone metastases were 75% at treated sites. Opioids were used by 72%, BMA was 27%, and chemotherapy was 50%. Re-irradiation rate was 22%. With a median follow-up of 5.8 months, 113 patients were alive, 129 had died, and 20 were lost follow-up. The average NRS was reduced from 6.1 to 3.1 from pretreatment to 3 months later per evaluable 232 targets, resulting in a pain relief rate of 60%. Opioids use (P<0.001) and re-irradiation (P<0.001) were significant factors of poor response in multivariate analysis. In our classification model, 89 targets were classified as class 1 (no opioids & no irradiation history), 211 were class 2 (other than class 1 and 3), and 66 were class 3 (opioids & re-irradiation), were 75/61/36% (P<0.001) of response rate. The 6-month cumulative pain progression rate was 12/22/32% (P<0.001), respectively. Palliative radiotherapy is highly effective for patients with painful tumors especially for those without previous irradiation history or use of opioids.