Introduction Home parenteral nutrition (HPN) for palliative malignancy is increasing and optimal selection of patients remains challenging. A nomogram predicting survival length1 based on Glasgow prognostic score (CRP and albumin), primary cancer, metastases & Karnofski performance status has recently been developed. In the nomogram survival was calculated from discharge date, rather than the date of PN commenced. We aimed to test the validity of the nomogram by retrospectively applying it to a cohort of HPN palliative cancer patients at St Mark’s Hospital & University Hospital Southampton.
Methods Adult patients with palliative cancer started on HPN were identified between 1/1/05 and 1/12/15. Patients were excluded with a diagnosis of psuedomyxoma or if a cancer developed while on HPN. Karnofski performance status was assigned though MDT discussion.
Results 47 patients met the inclusion criteria, 44 patients had sufficient data to complete the nomogram. There were 9 ovarian, 26 GI and 9 ‘other’ cancers (metastasis in 77%). Median survival for ovarian cancers was 100 days, GI cancers 55 days & other primaries 15 days. The nomogram over or underestimated survival length by ≤25%, 25–50%, ≥50% in 40%, 20% and 40% of patients, respectively. 12 patients were predicted to live <2 months, which was correct for 8 (67%). The mean overall difference in survival compared the nomogram prediction was 66 +/-59 days (95% CI). Figure 1, shows Kaplan-Meier curves for actual vs. predicted survival (log rank (Mantel-Cox) test p = 0.006) confirming significant difference.
Conclusion This data does not validate the use of this nomogram to predict survival length in HPN palliative malignancy patients. In our cohort the data shows that this nomogram most often under estimates survival. The retrospective nature of this study and relatively small number of patients is a limiting factor. Further prospective studies are needed and it is recommended defining survival as measured from date PN commenced. At present the best judge for considering palliative HPN is clinical acumen & patient performance status.
Reference 1 Bozzetti F, Cotogni P, Lo Vullo S, Pironi L, Giardiello D, Mariani L. Development and validation of a nomogram to predict survival in incurable cachectic cancer patients on home parenteral nutrition. Ann Oncol 2015;26(11):2335–2340.
Disclosure of Interest None Declared