Explainable Machine Learning Model to Predict Overall Survival in Patients Treated With Palliative Radiotherapy for Bone Metastases.

PURPOSE: The estimation of prognosis and life expectancy is critical in the care of patients with advanced cancer. To aid clinical decision making, we build a prognostic strategy combining a machine learning (ML) model with explainable artificial intelligence to predict 1-year survival after palliative radiotherapy (RT) for bone metastasis. MATERIALS AND METHODS: Data collected in the multicentric PRAIS trial were extracted for 574 eligible adults diagnosed with metastatic cancer. The primary end point was the overall survival (OS) at 1 year (1-year OS) after the start of RT. Candidate covariate predictors consisted of 13 clinical and tumor-related pre-RT patient characteristics, seven dosimetric and treatment-related variables, and 45 pre-RT laboratory variables. ML models were developed and internally validated using the Python package. The effectiveness of each model was evaluated in terms of discrimination. A Shapley Additive Explanations (SHAP) explainability analysis to infer the global and local feature importance and to understand the reasons for correct and misclassified predictions was performed. RESULTS: The best-performing model for the classification of 1-year OS was the extreme gradient boosting algorithm, with AUC and F1-score values equal to 0.805 and 0.802, respectively. The SHAP technique revealed that higher chance of 1-year survival is associated with low values of interleukin-8, higher values of hemoglobin and lymphocyte count, and the nonuse of steroids. CONCLUSION: An explainable ML approach can provide a reliable prediction of 1-year survival after RT in patients with advanced cancer. The implementation of SHAP analysis provides an intelligible explanation of individualized risk prediction, enabling oncologists to identify the best strategy for patient stratification and treatment selection.

Parenteral Nutrition in Palliative Cancer Care: Detrimental, Futile, or Beneficial?

Palliative cancer care patients may live for a long time, but malnutrition worsens the prognosis. Parenteral nutrition (PN) is suitable for replenishing a calorie deficit, but its advantages and tolerance late in the cancer trajectory are debated. We examined symptom development in hospitalized patients with and without PN. A total of 21 palliative cancer care patients receiving PN and 155 palliative cancer care patients not receiving PN during hospitalization in a specialized unit were retrospectively compared. We studied symptom intensity at admission, symptom relief during the hospital stay, and survival. The patients had locally advanced or metastatic cancer, a mean age of 70 years, and their median ECOG performance status was III. Symptom burden at admission was similar in the compared groups. Symptom relief during hospitalization was also similar. However, patients already on PN at admission reported more nausea and patients receiving PN during hospitalization reported better nausea relief compared to patients not receiving this intervention. Overall median survival was less than two months and similar in the compared groups. Based on a limited number of observations and a suboptimal study design, we were not able to demonstrate an increased symptom burden for palliative cancer care patients receiving PN late in the disease trajectory.

Inflammatory Markers and Radiotherapy Response in Patients With Painful Bone Metastases.

CONTEXT: Inflammation is proposed to influence tumor response in radiotherapy (RT). Clinical studies to investigate the relationship between inflammatory markers and RT response is warranted to understand the variable RT efficacy in patients with painful bone metastases. OBJECTIVES: To evaluate the association between inflammatory markers and analgesic response to RT in patients with painful bone metastases. METHODS: Adult patients from 7 European study sites undergoing RT for painful bone metastases were included in this prospective and longitudinal analysis. The association between RT response and 17 inflammatory markers at baseline, as well as the association between RT response and the changes observed in inflammatory markers between baseline and three and eight weeks after RT, was analyzed with univariate regression analyses. Baseline analyses were adjusted for potential clinical predictors of RT response. RESULTS: None of the inflammatory markers were significantly associated with an upcoming RT response in the analysis of 448 patients with complete baseline data. In patients available for follow-up, the three-week change in TNF (P 0.017), IL-8 (P 0.028), IP-10 (P 0.032), eotaxin (P 0.043), G-CSF (P 0.033) and MCP-1 (P 0.002) were positively associated with RT response, while the three-week change in CRP (P 0.006) was negatively associated. CONCLUSION: Results from this study show an association between RT response and change in pro-inflammatory mediators and indicate that inflammation may be important to achieve an analgesic RT response in patients with painful bone metastases. None of the investigated inflammatory markers were found to be pre-treatment predictors of RT response.

Benefits of Study Participation for Patients with Advanced Cancer Receiving Radiotherapy: A Prospective Observational Study.

Background: Patients with advanced cancer and bone metastases may have unmet palliative care (PC) needs that go unnoticed during clinical oncological practice. This observational study describes interventions that were initiated as the patients participated in the Palliative Radiotherapy and Inflammation Study (PRAIS). It was hypothesized that the patients would benefit from study participation due to PC interventions initiated by the study team. Methods: A retrospective review of patients' electronic records. Patients with advanced cancer and painful bone metastases included in PRAIS were eligible. All patients met with the study team before start of radiotherapy, after completion of Patient Reported Outcome Measures. Interventions initiated by the study team were documented in the patients' electronic records. Results: A total of 133 patients were reviewed: 63% males, mean (standard deviation [SD]) age 65 (9.6) and mean (SD) Karnofsky performance status (KPS) score 73.2 (9.1). Interventions were initiated in 50% (n = 67) of the patients. Changes in opioid management (69%), treatment of constipation (43%), and nausea (24%) and nutritional advice were most frequent (21%). Patients receiving interventions had lower mean KPS (70 vs. 77 p < 0.001), shorter survival time after study inclusion (median 28 vs. 57.5 weeks p = 0.005) and were more often opioid naïve (12% vs. 39% p < 0.001) than those not receiving interventions by the study team. Conclusions: Patients with advanced cancer and painful bone metastasis benefited from study participation due to multiple PC interventions initiated by the study team. The findings call for a systematic integration of PC in patients with advanced cancer. Trial Registration: ClinicalTrials.gov NCT02107664.

Clinical Predictors for Analgesic Response to Radiotherapy in Patients with Painful Bone Metastases.

BACKGROUND: Radiotherapy (RT) reduces pain in about 60% of patients with painful bone metastases, leaving many patients without clinical benefit. This study assesses predictors for RT effectiveness in patients with painful bone metastases. MATERIALS AND METHODS: We included adult patients receiving RT for painful bone metastases in a multicenter, multinational longitudinal observational study. Pain response within 8 weeks was defined as ≥2-point decrease on a 0-10 pain score scale, without increase in analgesics; or a decrease in analgesics of ≥25% without increase in pain score. Potential predictors were related to patient demographics, RT administration, pain characteristics, tumor characteristics, depression and inflammation (C-reactive protein [CRP]). Multivariate logistic regression analysis with multiple imputation of missing data were applied to identify predictors of RT response. RESULTS: Of 513 eligible patients, 460 patients (90 %) were included in the regression model. 224 patients (44%, 95% confidence interval (CI) 39%-48%) responded to RT. Better Karnofsky performance status (Odds ratio (OR) 1.39, CI 1.15-1.68), breast cancer (OR 2.54, CI 1.12-5.73), prostate cancer (OR 2.83, CI 1.27-6.33) and soft tissue expansion (OR 2.00, CI 1.23-3.25) predicted RT response. Corticosteroids were a negative predictor (OR 0.57, CI 0.37-0.88). Single and multiple fraction RT had similar response. The discriminative ability of the model was moderate; C-statistic 0.69. CONCLUSION: This study supports previous findings that better performance status and type of cancer diagnosis predicts analgesic RT response, and new data showing that soft tissue expansion predicts RT response and that corticosteroids is a negative predictor for RT response in patients with painful bone metastases.

Which factors can aid clinicians to identify a risk of pain during the following month in patients with bone metastases? A longitudinal analyses.

PURPOSE: Explore clinical factors associated with higher pain intensity and future pain in patients with bone metastases to identify patients who can benefit from closer follow-up or pain-modifying interventions. METHODS: This is a secondary analysis of 606 patients with bone metastases included in a multicenter longitudinal study. The dependent variables were "average pain" and "worst pain" in the last 24 h (0-10 NRS). Twenty independent variables with potential association to pain intensity were selected based on previous literature. Cross-sectional analyses were performed with multiple linear regression to explore factors associated with pain intensity at baseline. Longitudinal data were analyzed with a generalized equation models to explore current factors associated with pain intensity at the next visit in 1 month. RESULTS: Current pain intensity (p < 0.001), sleep disturbances (p 0.01 and 0.006), drowsiness (p 0.003 and 0.033) and male gender (p 0.045 and 0.001) were associated with higher average and worst pain intensity in 1 month. In addition, breakthrough pain was related to higher worst pain intensity (p 0.003) in 1 month. The same variables were also associated with higher average pain intensity at baseline. CONCLUSION: Higher current pain intensity, sleep disturbances, drowsiness, male gender, and breakthrough pain are factors associated with higher pain intensity in patients with bone metastases at the next follow-up in 1 month. These factors should be assessed in clinical practice and may aid clinicians in identifying patients that can benefit from closer follow-up or interventions to prevent lack of future pain control. TRIAL REGISTRATION IN CLINICALTRIALS.GOV : NCT01362816.

The Palliative Radiotherapy and Inflammation Study (PRAIS) - protocol for a longitudinal observational multicenter study on patients with cancer induced bone pain.

BACKGROUND: Radiation therapy (RT) results in pain relief for about 6 of 10 patients with cancer induced bone pain (CIBP) caused by bone metastases. The high number of non-responders, the long median time from RT to pain response and the risk of adverse effects, makes it important to determine predictors of treatment response. Clinical features such as cancer type, performance status and pain intensity, and biomarkers for osteoclast activity are proposed as predictors of response to RT. However, results are inconsistent and there is a need for better predictors of RT response. A similar argument can be stated for the development of cachexia; there are currently no predictors that can identify patients who will develop cachexia later in the cancer disease trajectory. Experimental and preclinical studies show that pain, depression and cachexia are related to inflammation. However, it is not known if inflammatory biomarkers can predict CIBP, depression or development of cachexia. METHODS: This multicenter, multinational longitudinal observational study will include 600 adult patients receiving RT for CIBP. Demographic data, clinical variables, osteoclast and inflammatory biomarkers will be assessed before start of RT, and 3, 8, 16, 24 and 52 weeks after last course of RT. The primary aim of the study is to identify potential predictors for pain relief from RT. Secondary aims are to explore potential predictors for development of cachexia, the longitudinal relationship between pain intensity and depression, and if inflammatory biomarkers are associated with changes in pain intensity, cachexia and depression during one-year follow up. DISCUSSION: The immediate clinical implication of the PRAIS study is to identify potential predictive factors for a RT response on CIBP, and thereby reduce non-efficacious RT. Patient benefits are fewer hospital visits, reduced risk of adverse effects and more individualized pain treatment. The long-term clinical implication of the PRAIS study is to improve the knowledge about inflammation in relation to CIBP, cachexia and depression and potentially identify associations and mechanisms that can be targeted for treatment. TRIAL REGISTRATION: ClinicalTrials.gov NCT02107664 , date of registration April 8, 2014 (retrospectively registered). TRIAL SPONSOR: The European Palliative Care Research Centre (PRC), Department of Clinical and Molecular Medicine, NTNU, Faculty of medicine and Health Sciences, Trondheim, N-7491, Norway.

Clinical and genetic factors related to cancer-induced bone pain and bone pain relief.

OBJECTIVE: The study objective was to evaluate whether there are clinical or genetic differences between patients with cancer-induced bone pain (CIBP) and patients with non-CIBP, and, in the CIBP group, in those with good versus poor opioid response. MATERIALS AND METHODS: A total of 2,294 adult patients with cancer who were receiving opioids for moderate or severe pain were included in the European Pharmacogenetic Opioid Study. Pain intensity and pain relief were measured using the Brief Pain Inventory. Linkage disequilibrium of 112 single nucleotide polymorphisms was evaluated in 25 candidate genes, and 43 haplotypes were assessed. Correlations among demographical factors, disease-related factors, genetic factors, CIBP, and pain relief were analyzed by logistic regression models corrected for multiple testing. Patients with bone metastases and bone/soft tissue pain were defined as having prevalent bone pain (CIBP population). This population was compared with patients who had other types of cancer pain (non-CIBP). RESULTS: A total of 577 patients (26.2%) had CIBP, and 1,624 patients (73.8%) had non-CIBP. Patients with CIBP had more breakthrough cancer pain episodes (64.2% vs. 56.4%, p = .001), had significantly higher pain interference in "walking ability in the past 24 hours" (p < .0001), used more adjuvant drugs (84.1% vs. 78.3%, p = .003), and had a higher, albeit nonsignificant, median overall survival (3.8 vs. 2.9 months, p = .716) than patients with non-CIBP. None of the examined haplotypes exceeded p values corrected for multiple testing for the investigated outcomes. CONCLUSION: Patients with CIBP who were taking opioids had a clinical profile slightly different from that of the non-CIBP group. However, no specific genetic pattern emerged for CIBP versus non-CIBP or for responsive versus nonresponsive patients with CIBP.