Nomagram reliability for predicting survival in patients with incurable cancer referred for home parental nutrition.
No Thumbnail Available
Authors
Cook, Hannah
Issue Date
2021
Type
Language
Keywords
Alternative Title
Abstract
"Introduction: In the presence of bowel obstruction, enterocutaneous fistula, short bowel, or severe mucosal disease, patients with incurable cancer are increasingly being referred for consideration of home parenteral nutrition (HPN). The decision to commence such treatment largely relies on expert opinion rather than robust data showing survival outcome. To address this shortcoming, a nomogram to predict median survival length in palliative cancer patients treated with HPN has been developed and validated.1 The nomogram is based on Glasgow prognostic score (CRP & albumin), primary cancer, metastases and Karnofsky performance status. The aim of this study was to assess the reliability and clinical value of the nomogram.
Methods: The nomogram was applied ambidirectionally to adult patients referred for palliative HPN between 1/3/15 and 7/7/20 at one tertiary HPN centre. Patients receiving chemotherapy or radiotherapy at the point of referral or during HPN treatment, and patients with neuroendocrine tumours were excluded. Intraclass correlation coefficient (ICC) was used to measure the reliability of the nomogram.
Results: 35 patients were identified. Eight patients were excluded due to commencing chemotherapy. Of the 27 remaining patients, 15 (66%) were female. 16 (59%) patients had primary GI cancers, six (22%) ovarian, and five (19%) other forms of cancer. Overall mean survival was 114 days (22-433) versus 104 days (30-200) for predicted survival (p=0.746). The nomogram over predicted survival in 59% of cases and under predicted in 33%. The predictions for seven patients (26%) were within 20% of their actual survival, 12 patients (44%) were within 50%, and the remaining patients between 50 and 248%. The ICC was 0.327 with a confident interval of -0.64-0.627, indicative of poor reliability.2
Conclusion: Although the p value suggests no significant difference between predicted and actual survival length, our study is limited by the small sample size. We considered a 20% variance between predicted and actual survival clinically acceptable; only a quarter of patients were within this range. Our study therefore does not support the use of the nomogram to predict survival in patients referred for palliative HPN and we should continue to use clinical acumen when considering such treatment. Further multi-centre research with larger sample sizes is needed before applying the nomogram to clinical practice."
Description
Citation
Kirk, C., Leyland, H., Thompson, N. et al. (2021) PTH-8?Nomogram reliability for predicting survival in patients with incurable cancer referred for home parenteral nutrition. Gut, 70 : A154-A155.
Publisher
Gut
License
Journal
Volume
Issue
PubMed ID
ISSN
1468-3288