# I-PREDICT

Source: https://onco.cc/trials/i-predict/  
OnCo record `i-predict` (Trial). Data CC BY-NC 4.0, attribute "Data from OnCo (onco.cc)"; commercial use needs a licence.

## TL;DR

I-PREDICT flipped the basket idea: instead of one drug for one mutation, a tumour board built a personal combination for each patient covering as many of their tumour's changes as possible. People whose treatment matched more of their alterations lived longer without progression.

## Summary

I-PREDICT was run from the UC San Diego Moores Cancer Center by Razelle Kurzrock and Jason Sicklick with a partner site at Avera Cancer Institute, opening in 2015. Patients with lethal, treatment-refractory cancers had tissue sequencing, circulating tumour DNA testing, PD-L1 and tumour mutational burden measured; a molecular tumour board then proposed a regimen, most often a combination of targeted drugs, immunotherapy and, where needed, chemotherapy, aimed at the largest possible share of each tumour's alterations. The study measured a matching score, the fraction of the tumour's characterised alterations addressed by the regimen.

The main report, in Nature Medicine in 2019, found that customised combinations were feasible in about half of the profiled patients, that toxicity was manageable with dose adjustment, and that patients with a matching score above 50 percent had higher disease control and longer progression-free and overall survival than those with lower scores. That association between degree of matching and outcome, rather than any single drug's result, was the study's contribution.

I-PREDICT influenced how molecular tumour boards think, shifting from one actionable alteration to the whole profile, and prompted follow-on studies of the same approach in newly diagnosed lethal cancers. Its limits are those of any single-arm study: patients with more alterations covered may also have had different biology, and only a randomised trial such as CUPISCO or ROME can show that matching itself causes the benefit.

## Fields

- Kind: Trial
- Status: completed
- Last checked: 2026-09-17
- Also known as: I-PREDICT study; Investigation of Profile-Related Evidence Determining Individualized Cancer Therapy
- Registry id: NCT02534675
- Phase: 2
- Setting: Treatment-refractory advanced solid tumours: tissue and blood genomic profiling reviewed by a molecular tumour board to build an individualised, usually combination, regimen for each patient, with outcomes analysed by how closely the regimen matched the tumour's alterations
- Sponsor: University of California, San Diego
- Result: Individualised combinations were feasible; a matching score above 50 percent was associated with better disease control and longer progression-free and overall survival.

## Sources

- ClinicalTrials.gov NCT02534675: https://clinicaltrials.gov/study/NCT02534675
- Sicklick et al., Nature Medicine 2019: molecular profiling of cancer patients enables personalized combination therapy, the I-PREDICT study: https://doi.org/10.1038/s41591-019-0407-5

## Connected records

- terms: [Basket, umbrella, and platform trials](https://onco.cc/terms/basket-umbrella-platform/), [Circulating tumour DNA (ctDNA)](https://onco.cc/terms/ctdna/), [Master protocol (platform, basket and umbrella trials)](https://onco.cc/terms/master-protocol/), [Rare cancers](https://onco.cc/terms/rare-cancers/), [Tumour mutational burden (TMB)](https://onco.cc/terms/tmb/)
- technologies: [Cancer variant knowledgebases and molecular tumour boards](https://onco.cc/technologies/variant-knowledgebases/), [Comprehensive genomic profiling](https://onco.cc/technologies/cgp/), [Liquid biopsy (ctDNA)](https://onco.cc/technologies/liquid-biopsy/)
- companies: [Foundation Medicine (Roche)](https://onco.cc/companies/foundation-medicine/)
- institutions: [UC San Diego Moores Cancer Center](https://onco.cc/institutions/ucsd-moores/)
- people: [Razelle Kurzrock](https://onco.cc/people/razelle-kurzrock/)
- bottlenecks: [Trial design, endpoints and cost](https://onco.cc/bottlenecks/b-trial-design/)

---
JSON: https://onco.cc/api/v1/entities/i-predict.json