The edition · Oncology
A renal cell prognostic model that beats IMDC, and AI colonoscopy's benefit turns out to be all diminutive polyps
CPI2, built on CheckMate-214 and validated on CheckMate-9ER, reached a C-index of 0.72 against 0.61 to 0.65 for IMDC. Separately, a network meta-analysis of 13 trials finds AI detection gains concentrate entirely in polyps of 5 mm or less.
The edition in brief
Four findings and a pearl for the oncology desk. CPI2, a new prognostic model for metastatic clear cell renal cell carcinoma, was developed on individual patient data from CheckMate-214 and externally validated on CheckMate-9ER. Built from 14 routinely available parameters - age, Karnofsky status, prior nephrectomy, four metastatic sites and seven laboratory values - it achieved discriminative accuracy at 36 months of 0.72 in both datasets, against 0.65 and 0.61 for IMDC in the same populations, with adequate calibration. PD-L1 expression added nothing. A network meta-analysis of 13 randomised trials and 4,156 patients found the benefit of AI-assisted colonoscopy is confined to diminutive polyps of 5 mm or less (standardised mean difference 0.21, 95% CI 0.07 to 0.35) with effects on small and large polyps of 0.02 and 0.01 - and even the diminutive finding carried a prediction interval from -1.12 to 1.54 and very low GRADE certainty. Eight randomised trials in 504 patients found exercise during neoadjuvant chemotherapy did not improve pathological complete response overall (risk ratio 1.08, 0.82 to 1.43), with an exploratory meta-regression signal confined to hormone-receptor-positive HER2-negative breast cancer. The practice-changer is a meta-analysis of nine randomised trials comparing early closure of a protective ileostomy within 30 days against late closure at 12 weeks or more after rectal cancer surgery: no difference in leak, morbidity, mortality, adjuvant chemotherapy delivery or low anterior resection syndrome, with faster return of flatus and lower costs. Plus a pearl on how a single temperature reading should be handled in a neutropenic patient.
CPI2 outperforms IMDC for metastatic clear cell renal cell carcinoma
Keep using IMDC to decide treatment for now, but expect CPI2 to appear in trial stratification - it discriminates substantially better on the same routinely available data.
AI colonoscopy finds more 5 mm polyps and essentially nothing else
Expect AI-assisted colonoscopy to add diminutive polyps and little else, and plan the downstream surveillance and pathology workload accordingly.
Exercise during neoadjuvant chemotherapy did not raise pathological complete response
Continue prescribing exercise during neoadjuvant chemotherapy for function and tolerance, and stop implying it improves tumour response.
One temperature of 38 degrees in a neutropenic patient is the whole diagnosis
In a neutropenic patient, one qualifying temperature means antibiotics within the hour - cultures are taken alongside, not before.
Closing a protective ileostomy within 30 days was as safe as waiting 12 weeks
In a selected patient with a sound anastomosis, offer ileostomy closure within 30 days rather than waiting 12 weeks - the trials show no difference in leak or morbidity and lower cost.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this one is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for oncology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free