Temporal Profiles of Symptom Scores after Palliative Radiotherapy for Bleeding Gastric Cancer with the Adjustment for the Palliative Prognostic Index: An Exploratory Analysis of a Multicenter Prospective Observational Study (JROSG 17-3)

Abstract

<h3>Purpose/Objective(s)</h3> Although palliative radiotherapy for gastric cancer may improve some symptoms, it may also have negative impact due to its toxicity. We investigated whether symptoms improved after radiotherapy with adjustment for the Palliative Prognostic Index (PPI) considering that patients with limited survival tend to experience deterioration of symptoms. <h3>Materials/Methods</h3> This study was an exploratory analysis of a multicenter prospective observational trial (Japanese Radiation Oncology Study Group [JROSG] 17-3). We assessed six symptom scores (nausea, lack of appetite, fatigue, shortness of breath, pain at the irradiated area, and distress) based on the M. D. Anderson Symptom Inventory at registration and 2, 4, and 8 weeks thereafter. PPI prediction model contains five independent indicators: palliative performance scale, intake, edema, dyspnea at rest, and delirium. A higher PPI score means worse survival. We tested whether symptoms linearly improved after adjusting for the baseline PPI. Shared parameter models were used to account for potential bias in missing data. <h3>Results</h3> The present study analyzed all 55 patients enrolled in JROSG 17-3. The median age was 73 years (range, 50–93 years). The average PPI was 2.9 (95% confidence interval: 2.2–3.6). The median 20 Gy (range, 8–45 Gy) total radiation dose was delivered in a median of 5 fractions. At registration, 27 (49%), 36 (65%), 44 (80%), 38 (69%), 29 (53%) and 44 patients (80%) had nausea, lack of appetite, fatigue, shortness of breath, pain, and distress, respectively. All six symptoms were evaluable for 48 (87%), 36 (65%), and 28 (51%) patients at 2, 4, and 8 weeks, respectively. With time from registration as the only explanatory variable in the model, a significant linear decrease was observed in shortness of breath, pain, and distress (slopes, −0.26, −0.22, and −0.19, respectively). Given that the interaction terms (i.e., PPI × time) were not significantly associated with symptom scores in any of the six symptoms, only PPI was included as the main effect in the final multivariable models. After adjusting for the PPI, shortness of breath, pain, and distress significantly improved (slope, −0.25, −0.19, and −0.17; p < 0.001, 0.002, and 0.047, respectively). Higher PPI scores were associated with higher symptom scores in all six symptoms. <h3>Conclusion</h3> Shortness of breath, pain, and distress significantly improved after radiotherapy. Moreover, a higher PPI was significantly associated with higher symptom scores at all time points, including baseline. In contrast, PPI did not seem to influence the improvement of these symptoms. Regardless of the expected survival, patients receiving radiotherapy for gastric cancer can expect improvement in shortness of breath, pain, and distress over 8 weeks.