DIATHERA^® · OVARIAN CANCER RESEARCH

Investigating treatment benefitfrom the H&E slide in ovarian cancer.

Name: Investigating treatment benefit from the H&E slide in ovarian cancer.
Brand: DiaDeep

Which patients benefit most from maintenance therapy in advanced ovarian cancer.

RESEARCH USE ONLY Not for diagnostic or treatment-selection use

Help match treatment intensity to each patient’s needs.

Recurrence risk and expected treatment benefit are central to avoiding overtreatment and undertreatment. These questions drive the development of DiaSurv^®^* and DiaThera^®^*.

Treatment benefit research in advanced high-grade ovarian cancer, read from the pretreatment slide.

0new
biopsy.

DiaThera^® reads the pretreatment H&E slide already taken.

0extra
tissue.

No additional sampling for the analysis.

0new test
to order.

It uses the H&E slide and the HRD status already in the record.

TWO RESEARCH QUESTIONS

Two studies for ovarian cancer
treatment response prediction.

One study evaluates heterogeneity of benefit from a specific maintenance regimen. A separate study evaluates whether HRD status can be predicted from routine H&E slides.

Treatment benefit

In PAOLA‑1, a treatment-aware model compared predicted progression risk under olaparib plus bevacizumab and bevacizumab alone to derive a continuous ETI score.

HRD status

A separate model evaluated whether HRD status could be predicted from routine H&E slides. This work investigates support or triage for molecular testing, not its replacement.

421 patients with pretreatment slides in the PAOLA‑1 treatment-benefit analysis.

0.36 treatment-by-ETI interaction HR for PFS‑1 (95% CI 0.22–0.59).

2 studies peer-reviewed ovarian cancer studies covering treatment benefit and HRD status.

PAOLA‑1 · TREATMENT-EFFECT CONTRAST

DiaThera separated patients by estimated benefit from adding olaparib.

At 36 months, the estimated improvement in the probability of remaining progression-free with olaparib plus bevacizumab versus bevacizumab alone was 36.0 percentage points greater in the DiaThera-predicted responder group than in the non-responder group.

36.0 pp DiaThera-predicted groups 95% CI 17.7 to 54.2

35.8 pp HRD-defined groups 95% CI 18.9 to 54.2

DiaThera

36.0

HRD

35.8

For each method, the chart compares estimated treatment benefit in the positive or responder group with the negative or non-responder group. The point marks the estimate, the bar its 95% confidence interval, and each dot one percentage point. Source: Frenel J-S, et al. npj Digit Med. 2026.

HRD STATUS · ETI STRATIFICATION

Related signals.
Not the same classification.

ETI stratified predicted treatment benefit within both HRD-positive and HRD-negative groups in the 421-patient PAOLA‑1 analysis.

HRD-positive 243

195ETI responder
80.3%

48ETI non-responder
19.7%

HRD-negative 178

6ETI responder
3.4%

172ETI non-responder
96.6%

ETI responder status was defined in the published analysis from the model-derived treatment-effect score. Source: Frenel J-S, et al. npj Digit Med. 2026.

H&E-BASED HRD PREDICTION

Developed once.
Evaluated across cohorts.

A separate deep-learning study evaluated HRD prediction from routine whole-slide images in one development cohort and two external cohorts.

Development 0.73 AUC

IHP-OV

257 patients

External evaluation 0.70 AUC

PAOLA‑1

468 patients · AUC 0.72 in BRCA-wildtype

External evaluation 0.69 AUC

TCGA

89 patients

NEGATIVE PREDICTIVE VALUE

Up to 0.98 for identifying HRD-negative cases in the reported analyses.

PROGRESSION-FREE SURVIVAL

In olaparib-treated BRCA-wildtype patients, AI-predicted negative groups showed progression-free survival curves nearly identical to Myriad-defined negative groups.

Source: Frenel J-S, et al. Int J Gynecol Cancer. 2026;36:104717. HRD prediction and treatment-benefit estimation are separate research models and analyses.

RESEARCH COLLABORATION

Turn treatment outcomes into a research program.

Bring together whole-slide images, treatment assignments and longitudinal outcomes with the DiaDeep team.

Discuss a research project

PEER-REVIEWED EVIDENCE

Read the
source studies.

Review the populations, endpoints, methods and limitations reported in the original ovarian cancer publications.

2026

Treatment-aware deep learning enables counterfactual prediction of individual benefit from PARP inhibitors in ovarian cancer

Homologous recombination deficiency (HRD) assays are used to select patients with ovarian cancer for PARP inhibitors, but they do not fully capture the…

Deep learning-based prediction of homologous recombination deficiency from histopathological whole-slide images in ovarian cancer

Homologous recombination deficiency is a critical biomarker for guiding first-line treatment with poly(adenosine diphosphate-ribose) polymerase inhibitors in…

Artificial intelligence to predict homologous recombination deficiency in ovarian cancer from whole-slide histopathological images.

In the field of ovarian cancer diagnosis, predicting Homologous Recombination Deficiency (HRD) holds paramount importance for personalized treatment…