Precision Oncology/Recurrence Calculator
The study genotic.com

Clear-cell RCC · post-nephrectomy recurrence · research tool

Which resected kidney cancers will recur?

The Leibovich score guides adjuvant-therapy decisions after nephrectomy for clear-cell renal cell carcinoma — but most patients who go on to recur sit in its low / intermediate range and are never selected for treatment. This tool adds two tumour biomarkers to the Leibovich score — a systemic inflammatory index (PIV with the CD4/CD8 ratio) and BAF60b — to sharpen that decision and, above all, to flag the recurrences the score misses. It returns an individual 5-year recurrence probability; enter the pathology and three markers below.

0.84AUC
Cross-validated discrimination of the full model, vs 0.76 for the Leibovich score alone.
33/ 38
Future recurrences flagged among patients Leibovich calls low/intermediate risk (87 %, cross-validated).
n = 109
Development cohort — 60 recurrence-free vs 49 with 5-year recurrence. Single-centre; external validation pending.

Try a worked example

A patient the Leibovich score calls low-risk (score 2) — pT1b, node-negative, <10 cm, no necrosis, grade 2 — but with high inflammation and BAF60b (PIV 620, CD4 45 %, CD8 22 %, BAF60b 95 / 140). The model flags a markedly higher risk.

1

Leibovich score

Compute from pathology, or enter a known score.

Primary tumour stage (pT) Pathological primary-tumour stage from the nephrectomy specimen (AJCC/TNM). pT1a ≤4 cm · pT1b 4–7 cm · pT2 >7 cm confined to kidney · pT3–pT4 beyond the kidney.
Regional lymph nodes Regional lymph-node status. pNx = not assessed · pN0 = negative · pN1/pN2 = tumour in one or more regional nodes.
Tumour size Greatest tumour diameter on the pathology report. The Leibovich score uses a single 10 cm cut-off.
Necrosis Presence of histological (microscopic) coagulative tumour necrosis in the specimen, as reported by pathology.
Nuclear grade Nuclear grade from the pathology report (Fuhrman or WHO/ISUP). Grade 1–2 low; grade 3 and grade 4 carry increasing points.
–Leibovich score
2

Inflammatory markers & BAF60b

PIV from the blood count; CD4/CD8 and BAF60b from tumour IHC.

PIV — pan-immune-inflammation value From a routine pre-operative full blood count: neutrophils × monocytes × platelets ÷ lymphocytes (absolute counts). Reflects systemic inflammation.
CD4 % of TILs Percentage of tumour-infiltrating lymphocytes positive for CD4 by immunohistochemistry, counted over representative high-power fields.
CD8 % of TILs Percentage of tumour-infiltrating lymphocytes positive for CD8 by immunohistochemistry. In RCC, higher CD8 / lower CD4:CD8 is adverse.
BAF60b — nucleus H-score BAF60b (SMARCD1) IHC H-score of nuclear staining in tumour cells. H-score = Σ(intensity 0–3 × % cells) = 0–300.
BAF60b — cytoplasm H-score BAF60b (SMARCD1) IHC H-score of cytoplasmic staining in tumour cells (0–300). The model uses nuclear + cytoplasmic total.

The estimate also updates live as you type.

Estimated 5-year risk

Enter the Leibovich score and all three biomarkers to see the estimated risk.

The model

index = PIV × (CD8 + 0.5) / (CD4 + 0.5)  ·  BAF = nucleus + cytoplasm H-score
logit(risk) = −6.570 + 0.573·Leibovich + 0.430·ln(index) + 0.715·ln(1 + BAF)

Multivariable logistic regression; all three factors independently significant (Leibovich OR 1.77, inflammatory index OR 1.54, BAF60b OR 2.04). Cross-validated AUC 0.84 versus 0.76 for the Leibovich score alone. See the full study report for methods and validation.

Research use only. This calculator implements a model derived from a single-centre cohort of 109 patients and has not been externally validated. It is intended to accompany the associated study and for research discussion — it is not a validated clinical decision aid, must not replace clinical judgment, multidisciplinary assessment or established guidelines, and no decision threshold has been established. The percentage is a modelled probability, not a directive.

About this tool — what is calculated, and by whom

What it calculates

An individual probability of recurrence within 5 years after nephrectomy for clear-cell renal cell carcinoma (ccRCC). The estimate combines the established Leibovich progression score with two tumour-derived signals that carry information the score does not: a systemic inflammatory index and BAF60b expression.

How it is calculated

A multivariable logistic-regression model. The inflammatory index is PIV × (CD8+0.5) / (CD4+0.5) — the pan-immune-inflammation value divided by the tumour CD4/CD8 ratio; BAF60b is the sum of nuclear and cytoplasmic H-scores. Each factor is entered on its natural or log scale and all three remain independently significant (odds ratios: Leibovich 1.77, inflammatory index 1.54, BAF60b 2.04).

How well it performs

Estimated by repeated stratified cross-validation (5-fold × 100). The full model reaches an AUC of 0.84 versus 0.76 for the Leibovich score alone (ΔAUC +0.08; 100 % of repeats favour the full model; Nadeau–Bengio p<0.001). Its main value is in the Leibovich low/intermediate group, where it flags 33 of 38 (87 %) future recurrences the score would miss.

Where the data come from

A single-centre development cohort of 109 patients (60 recurrence-free, 49 with 5-year recurrence). Markers were measured on the resected tumour (immunohistochemistry) and a pre-operative blood count. Results are internally cross-validated and await external, multi-centre validation before any clinical use.

Whose work this is

The model and this calculator were developed by Genotic (computational and statistical analysis) together with the collaborating clinical and pathology research team who assembled the cohort and biomarker data. Full methods, figures and validation are in the accompanying study report. Enquiries: greg@genotic.com.

How to read the result

The percentage is a modelled 5-year risk, shown alongside the standard Leibovich category. A discordance flag appears when Leibovich is reassuring (low/intermediate) but the model is high — the situation this tool is built to surface. No treatment threshold is defined; interpret within the full clinical picture.