DIAKWANT® · DIGITAL ASSAY · PAN‑ORGAN

DiaKwant® PD‑L1CPS and TPS

Review the tumor regions and cell counts behind CPS and TPS to check how the score was calculated.

Histology tissue at cellular scale
CE MARKED UNDER IVDR, CLASS C, IN THE EUResearch Use Only outside the EU

VALIDATED SCORING SCOPE

Pan-organ PD-L1 scoring.

CPS
Breast · Head and neck · Gastrointestinal · Cervix
TPS
Lung · Urothelial

HOW IT WORKS

From slide
to reviewable score.

The same three steps apply to every DiaKwant® assay; the analysis in the middle is specific to the biomarker.

INPUT

Whole-slide image

Access the slide for quantification directly through your supported image management system.

ANALYSIS · PAN‑ORGAN

PD‑L1

Review the tumor regions and cell counts behind CPS and TPS to check how the score was calculated.

REVIEW

Reviewable quantification result

The assay returns the score and supporting counts alongside the heatmap for expert review.

AVAILABLE ON

EVALUATION PROTOCOL

Tested against
expert consensus.

Every score (AI alone, pathologists alone, pathologists assisted by DiaKwant®) is compared with the same reference.

  • DiaKwant®AI alone
  • PathologistsWithout AI
  • Pathologists + DiaKwant®Same readers, after a washout phase

REFERENCE

Consensus score3 expert pathologists from 3 laboratories. Discordances are resolved by discussion.
  • 2studies
  • 291cases
  • 4tumour types
  • 3expert pathologists set the reference

Pathologists alonePathologists + DiaKwant®With DiaKwant® assistanceExpert consensus

ACCURACY

Out of 100 cases, 6 more correctly classified

77.6%Pathologists alone

83.9%With DiaKwant®

p < 0.01 · DiaKwant® alone: 83.1%

SENSITIVITY

Out of 100 positive cases, 11 more detected

78.5%Pathologists alone

89.6%With DiaKwant®

DiaKwant® alone: 91.2% · Specificity  73.4% → 76.2%

AGREEMENT BETWEEN PATHOLOGISTS · ICC

From moderate to excellent reliability

MODERATE 0.5–0.75GOOD 0.75–0.9EXCELLENT > 0.9
0.763Expert consensus0.705Pathologists alone0.904Pathologists + DiaKwant®

Each dot is one case out of 100. Boulagnon-Rombi C. et al., Impact de l’assistance par intelligence artificielle sur le scoring PD-L1, poster, Assises de Pathologie 2026 · ASCO 2026 · ECP 2026 · ECDP 2026. Bossard C. et al., Clinical evaluation of an automated pan-organ combined PD-L1 scoring using artificial intelligence on immunostained whole-slide images, ESMO Real World Data and Digital Oncology 2025.

PUBLICATIONS AND ABSTRACTS

Evidence for PD‑L1.

Studies and congress abstracts listing PD‑L1 as a biomarker. The full list is on the Resources page.

All publications ↗︎

FAQ

Practical
questions.

Practical information about PD‑L1 scoring with DiaKwant®.

01What does DiaKwant PD‑L1 calculate?

DiaKwant PD‑L1 calculates a Combined Positive Score (CPS) or a Tumor Proportion Score (TPS), depending on the organ being assessed.

02Which cancers use CPS and which use TPS?

CPS is used for breast, head and neck, gastrointestinal and cervical cancers. TPS is used for lung and urothelial cancers.

03Is the PD‑L1 analysis automated?

Yes. The quantitative analysis is automated: DiaKwant processes the whole-slide image and returns the calculated score.

04What information can the pathologist review?

The pathologist can review the cell counts behind the score and a heatmap showing where positive cells were identified on the slide.

05Does DiaKwant replace the pathologist?

No. DiaKwant is a diagnostic aid. The result must be reviewed by a qualified pathologist, who retains full responsibility for the final interpretation and report.

06How can a laboratory confirm technical compatibility?

Technical compatibility depends on your laboratory's scanner and image management configuration. Contact the DiaDeep team to confirm compatibility for your validated setup.

PD‑L1 ON YOUR SLIDES

See DiaKwant® PD‑L1 at work on your own anonymized slides.

Discuss your scoring context and validated technical configuration with the DiaDeep team.

Try it on your cases ↗︎Apply to Club Quanti ↗︎