Skip to content
PaperFren

Can a clinicopathologic model predict cSCC metastasis risk?

Open paper intelligence

A Dutch-developed, England-validated model (C-index ~0.80–0.84) gives personalised absolute metastatic risk for cutaneous squamous cell carcinoma better than current staging for key decisions.

Source

Personalised decision making to predict absolute metastatic risk in cutaneous squamous cell carcinoma: development and validation of a clinico-pathological model

Rentroia-Pacheco B, Tokez S, Bramer EM, et al. · EClinicalMedicine · 2023

doi.org/10.1016/j.eclinm.2023.102150Read the full paper ↗28 citationscc by

What they did

Built nested case–control cohorts from nationwide Dutch (2007–2008) and English (2013–2015) cSCC data, developed a weighted Cox clinico-pathological model (8/11 variables) on n=390, validated on n=696, and compared discrimination/calibration/net benefit with BWH and AJCC staging.

What they found

Optimism-corrected C-index 0.80 in development and 0.84 in validation; calibration slope 0.96; improved decision-curve net benefit for follow-up/adjuvant thresholds. Released as an online calculator.

The limits

What it doesn't show

Still needs prospective decision-impact studies in other populations; nested case–control design estimates relative/absolute risk under sampling weights, not a simple consecutive series prevalence.

Key terms

cSCC
Cutaneous squamous cell carcinoma of the skin.
Absolute metastatic risk
Personalised probability of metastasis rather than stage category alone.
C-index
Discrimination metric for ranking who has higher vs lower risk.
Calibration slope
Agreement between predicted and observed risks (near 1 is ideal).
Nested case–control
Cases with metastasis matched to non-metastatic controls within nationwide cohorts.
Decision curve analysis
Compares clinical net benefit across decision thresholds.

Flashcards

1 / 9

Research intelligence for this paper

See its role on concept claims, tensions it is part of, placement history, and related discoveries.

Open paper intelligence

Quiz yourself

1 / 5

Validation cohort size:

Common questions

Development vs validation N?

390 Dutch; 696 English.

Validation C-index?

0.84 (95% CI 0.81–0.87).

How many predictors kept?

Eight of eleven clinico-pathological variables.

Where is the tool?

Online web-based calculator linked in the paper.

More on Decision making