Abstract

Kantify reports a preclinical proof-of-concept program in soft tissue sarcoma with a focus on liposarcoma, performed in several steps:

• Kantify conducted target discovery and hit discovery in parallel using Kantify’s AI platform, Sapian, which yielded a list of novel drug candidates and novel drug targets.

• The first drug candidates and targets to be tested experimentally were selected by the team of Dr. Maya Jeitany at Nanyang Technological University and validated on in vitro assays.

• The list of candidates was independently validated on in vitro assays by Institut Bergonié under the leadership of Professor Antoine Italiano.

This workflow produced an in vivo-ready candidate.

• Institut Bergonié expanded the candidate portfolio (based on predictions from Sapian) through additional in vitro assessments and conducted in vivo validation.

• With support from the INCITE Foundation, the program advanced to show increased overall survival in cell line-derived xenograft models in mice. The next phase prioritizes rational combinations to achieve substantial and durable tumor control.

Problem Statement

Soft tissue sarcoma is a group of rare and heterogeneous malignancies with limited treatment options. Liposarcoma exemplifies these constraints, with few validated targets and frequent resistance.

Conventional discovery pipelines typically go through several steps: target prediction, target validation, and only then medicinal exploration of modulators.

This sequential design can unfortunately prolong timelines and create gaps between biological causality and chemical tractability. This problem is especially important in rare diseases such as liposarcoma. Novel approaches are therefore needed.

Methodological novelty in rare disease and rare cancer

Kantify employed a parallelized methodology that is novel in the context of rare diseases and rare cancers and relies on its AI platform Sapian.

Sapian simultaneously: * Predicted novel targets involved in liposarcoma. * Prioritized small-molecule candidates predicted to modulate these targets.

By aligning target nomination and candidate prediction from the outset, the program reduced the elapsed time from project start to in vitro candidate validation from a typical 24 months to 4 months. This parallelization maintained coherence between disease biology and chemistry while accelerating the learning cycle.

Collaboration and roles

Kantify and Nanyang Technological University (Dr. Maya Jeitany): joint execution of parallel target discovery and hit discovery using Sapian. Dr Jeitany performed target validation of the targets predicted by Sapian, and hit validation of the hits predicted by Sapian.

Institut Bergonié (Dr Tony Sourisseau): independently validated the candidates predicted by Sapian through in vitro and in vivo assays.

INCITE (Dr. JC Neel, PhD): financial support enabling expansion to additional candidates and in vivo studies.

Results

Initial validation from parallel discovery

The first in vivo-ready candidate, derived from the Sapian-enabled parallel track, produced a measurable increase in overall survival in liposarcoma CDX models compared to untreated control. The observed pharmacology was consistent with the originating target hypothesis.

Independent reproduction, portfolio expansion, and in vivo validation at Institut Bergonié

Institut Bergonié reproduced the core signal. Under the direction of Professor Antoine Italiano, the team evaluated additional molecules generated by the same discovery funnel and validated them in vitro, thereby expanding the candidate portfolio. Institut Bergonié also conducted in vivo validation. One candidate increased overall survival in mouse CDX models relative to untreated control.

Program expansion enabled by INCITE

Philanthropic support from INCITE allowed confirmatory work and broader candidate evaluation. This support enabled progression of the in vivo studies that demonstrated an increase in overall mouse survival.

In parallel, INCITE enabled Kantify to access the competencies of complementary biotech companies, such as Rosebud Biosciences, which performed off target toxicity testing on iPSC-derived organoids through its state-of-the-art robotics platform, and Vibe Bio, which collected and ranked the different IP and market positioning characteristics of Kantify’s repurposable candidates through its dedicated AI technology.

Discussion

A parallelized workflow that connects causal target hypotheses to tractable chemistry can compress timelines and improve translational coherence. In a setting with scarce data and limited treatment options such as soft tissue sarcoma, the approach provides a credible path to early PoC.

Acknowledgments

We thank Dr. Maya Jeitany and colleagues at Nanyang Technological University for collaborative discovery work in target and hit identification. We thank Professor Antoine Italiano and Tony Sourisseau at Institut Bergonié for independent reproduction, in vitro assessment of additional candidates, and in vivo validation. We are grateful to Dr. JC Neel, PhD and INCITE for support that enabled program expansion.

Disclosure

All results reported here are preclinical and non-GLP. No claims are made regarding safety or efficacy in humans.