Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-2114
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-21T18:13:07.912117Z
status: published
lens: g_space
sector: health
headline: Non-contrast CT radiomics for identifying nonfunctioning kidneys: a multicenter development and external validation study
dek: In urological practice, the decision to resect or preserve a kidney with obstruction, tumor or atrophy is based on split renal function, although SPECT renal dynamic scintigraphy as the reference standard is resource-intensive. We developed and externally validated a non-contrast CT radiomics model to identify nonfunctioning kidneys and support urological decision-making in this setting. In this three-center retrospective study (TRIPOD+AI), adults underwent non-contrast CT and SPECT renal dynamic scintigraphy wi…
gain_title: A radiomics machine learning model using routine non-contrast CT identified nonfunctioning kidneys with high negative predictive value, potentially reducing unnecessary SPECT scintigraphy and informing nephrectomy versus kidney-sparing decisions.
problem_title: (none)
trace_subject: (none)
gain_reading: A radiomics machine learning model using routine non-contrast CT identified nonfunctioning kidneys with high negative predictive value, potentially reducing unnecessary SPECT scintigraphy and informing nephrectomy versus kidney-sparing decisions.
gain_evidence: Non-contrast CT radiomics identified nonfunctioning kidneys without contrast agents or radiotracers
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers built machine learning models from radiomic features of non-contrast CT scans to identify kidneys that no longer function. Data came from three centers, with adults who had both a CT and a SPECT renal dynamic scintigraphy scan within 72 hours; SPECT served as the reference standard.

This matters because deciding whether to remove or preserve an obstructed, tumorous or atrophic kidney depends on split renal function, and SPECT is resource-intensive. The work is retrospective and the tool is framed as a rule-out aid, so its value in real clinical decisions still needs to be shown in practice.
limitation: The study is retrospective, and the advantage over the combined baseline was not statistically significant in one external cohort. External sensitivity of the random forest at fixed thresholds was lower than internally.
tag: Evidence-backed gain
key_points: The study followed TRIPOD+AI reporting and included 1029 patients across three centers, with 418 in the internal cohort. | The random forest model reached an internal test AUC of 0.9426, sensitivity of 85.5%, specificity of 88.3% and negative predictive value of 95.6%. | External AUCs for the random forest were 0.875 and 0.913, and for the nomogram 0.920 and 0.881. | The Rad-score correlated with single-kidney GFR (ρ = -0.738, P < 0.001) and helped with risk stratification.
rundown: The pipeline extracted IBSI-compliant radiomics features with PyRadiomics and built between-kidney asymmetry features. Features were narrowed through Spearman filtering, recursive feature elimination and LASSO, leaving 23 features, and 12 machine learning models were developed and tested in two independent external cohorts.

At thresholds fixed in the training data, the random forest reached external sensitivities of 66.0% and 76.4% with negative predictive values of 94.8% and 93.3%. The nomogram reached sensitivities of 92.5% and 93.7% with negative predictive values of 98.3% and 97.2%.

Both radiomics models outperformed logistic-regression baselines built from clinical variables and simple CT size and asymmetry measures. Non-contrast CT is already part of routine urological workup, so the approach needs no contrast agents or radiotracers.
sources:
- peer_reviewed | International Urology and Nephrology | https://doi.org/10.1007/s11255-026-05451-3 | 2026-10-20
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

a77580e7d051083b1421d16b47a667b4edd7ead219ea55f915567795cf288da3

previous

0000000000000000000000000000000000000000000000000000000000000000