Palliative Care

SS 16 - Palliative 1

95 - Pilot Assessment of the BMET Decision Support Platform: A Tool to Improve Provider Survival Estimates and Selection of Prognosis-Appropriate Treatment for Patients with Symptomatic Bone Metastases

Monday, September 16
4:15 PM - 4:25 PM
Location: Room W190

Pilot Assessment of the BMET Decision Support Platform: A Tool to Improve Provider Survival Estimates and Selection of Prognosis-Appropriate Treatment for Patients with Symptomatic Bone Metastases
S. R. Alcorn1, J. Fiksel2, C. Hu3, J. L. Wright1, L. R. Kleinberg1, A. S. Levin4, T. Smith5, Z. Cheng1, C. R. Elledge1, K. Kim1, A. D. Rao1, L. Sloan1, B. R. Page1, S. F. Stinson1, R. Voong1, T. R. McNutt1, M. R. Bowers1, T. L. DeWeese1, and S. Zeger2; 1Department of Radiation Oncology and Molecular Radiation Sciences, Johns Hopkins University School of Medicine, Baltimore, MD, 2Department of Biostatistics, Johns Hopkins School of Public Health, Baltimore, MD, 3The Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins University School of Medicine, Baltimore, MD, 4Department of Orthopaedic Surgery, Johns Hopkins University School of Medicine, Baltimore, MD, 5Department of Palliative Care, Johns Hopkins University School of Medicine, Baltimore, MD

Purpose/Objective(s): Previously, we developed the BMET random survival forests model, which uses 27 covariates to estimate survival in patients seen in consultation for palliative radiotherapy (RT) to symptomatic bone metastases. In the present study, we conducted a pilot assessment of the clinical utility of this model and an associated decision support platform (DSP).

Materials/Methods: To facilitate clinical use of the BMET model, we constructed a DSP that (1) collects covariate data and displays a patient-specific predicted survival curve, and (2) provides case-specific, evidence-based recommendations for RT, open surgery, systemic therapy (ST), and hospice referral in the care of symptomatic bone metastases. Five trainee and 5 attending physicians in Radiation Oncology participated in the DSP assessment. A total of 55 patient cases were randomly selected from the 397 patients used to build the BMET model; each predicted survival curve displayed as part of the DSP was refitted leaving the case patient out. Relevant case data including BMET covariates were summarized and presented to physicians at 2 times: without and then with use of the DSP (separated by a 3-week washout). At each time, physicians were asked to estimate patient survival in the 12 months following RT consultation; their confidence in and likelihood of sharing this estimate with the patient (increasing 1-10 scales); recommendations for open surgery, ST, and hospice referral; and preferred RT regimen (0, 1, 5, 10, or >10 conventional fractions or stereotactic RT). Wilcoxon signed-rank test evaluated paired survival estimates and rating scales, and McNemar’s test compared accuracy of survival estimates at clinically relevant binary time points.

Results: Assessment completion rate was 96%. Pre- vs. post-DSP, physicians' estimates of survival were mean 7.9 (SD 3.6) vs. 6.9 (SD 3.7) months, respectively, p<0.001. There was a significant reduction in overestimation of true minus estimated survival time, with a mean difference of -2.1 (SD 4.1) vs. -1 month (SD 3.5), p<0.001. This improvement was observed across training level. Accuracy of survival predictions were significantly improved at binary time points of <3 (72 vs. 79%, p<0.001), ≤6 (64 vs. 71%, p=0.007), and ≥12 months (70 vs. 81%, p<0.001). Median ratings of confidence in and likelihood of sharing prognosis each increased from 6 to 8, both p<0.001. There was greater concordance in matching use of 1-fraction RT with true survival <3 months (70 vs. 76%, p=0.001) and <10 fraction RT with true survival <12 months (55 vs. 62%, p=0.006). Improvements noted in hospice referral, ST, and surgical interventions were not statistically significant.

Conclusion: In this pilot study, use of the BMET-DSP significantly improved physician accuracy in estimating survival and increased prognostic confidence, likelihood of sharing prognosis, and use of prognosis-appropriate RT regimens in the care of symptomatic bone metastases. These data support future multi-institutional validation of the DSP.

Author Disclosure: S.R. Alcorn: Employee; Johns Hopkins University. Research Grant; Elekta AB, National Institutes of Health. Committee member; ASTRO. J. Fiksel: Research Grant; National Institute on Aging. C. Hu: Employee; Food and Drug Administration. Research Grant; National Cancer Institute. Consultant; Merck & Co. J.L. Wright: Honoraria; ASTRO. L.R. Kleinberg: Research Grant; Accuray, Novocure, Arbor. Consultant; Accuray, Novocure. Travel Expenses; Accuray. A.S. Levin: None. T. Smith: Fellow; ASCO, American Association of Hospice and Palliative Medicine. Z. Cheng: None. C.R. Elledge: None. K. Kim: None. A.D. Rao: Employee; Johns Hopkins University. Research Grant; Elekta AB. L. Sloan: Employee; Johns Hopkins Universitty. B.R. Page: CHEDI committee; ASTRO. Education Committee; ASTRO. R. Voong: Research Grant; Radiation Oncology Institute, Lung Cancer Research Foundation, Canon, Inc. T.R. McNutt: Research Grant; Elekta Oncology Systems, Philips Radiation Oncology Systems, Canon, Radiation Oncology Institute. Stock; Oncospace Inc. Patent/License Fees/Copyright; Accuray-Tomotherapy, Sun Nuclear. President; AAPM-MAC. Chairman of the Board; Oncospace Inc. M.R. Bowers: None. T.L. DeWeese: President; ASTRO. S. Zeger: Honoraria; ASTRO.

Sara Alcorn, MD, PhD, MPH

Johns Hopkins University

Disclosure:
Employment
Family: Johns Hopkins University: Assistant Profesor: Employee; Johns Hopkins University: Assistant Professor: Employee

Compensation
Elekta AB: Research Grants

Leadership
ASTRO: Committee member: Committee on Health Equity, Diversity and Inclusion;
ASTRO: Committee member: Research Grants Evaluation Subcommittee

Presentation(s):

Send Email for Sara Alcorn


Assets

95 - Pilot Assessment of the BMET Decision Support Platform: A Tool to Improve Provider Survival Estimates and Selection of Prognosis-Appropriate Treatment for Patients with Symptomatic Bone Metastases



Attendees who have favorited this

Please enter your access key

The asset you are trying to access is locked. Please enter your access key to unlock.

Send Email for Pilot Assessment of the BMET Decision Support Platform: A Tool to Improve Provider Survival Estimates and Selection of Prognosis-Appropriate Treatment for Patients with Symptomatic Bone Metastases