What does the generative AI add compared with NTT Data's existing service?

The earlier service, launched in July, uses SQL, Python and other programming languages to extract candidates. The generative AI adds written reasons for exclusion and suggestions such as when a patient might become eligible.

How much did the earlier screening cut the number of patients doctors had to review?

In a technical validation with Kindai University Hospital on lung-cancer trials, about 3,000 patients' data was analyzed. The number of patients needing detailed review fell by 97.4% and 99.4% in two trials.

What is NTT Data's revenue target for this business?

NTT Data aims for sales of 1 billion yen or more in fiscal 2029, combining this service with generative-AI-based services.

Large Language ModelsAI in HealthcareTop Companies' AI MovesTop Companies AI12h agoNov 14, 2026, 06:31 JST

NTT Data, 中外製薬 test tsuzumi2 for trial screening

Send on LINE

3 Key Points

What happened
NTT Data said it is working with Chugai Pharmaceutical and Kindai University Hospital on joint research that matches trial criteria against electronic medical records using a medical version of NTT's tsuzumi2 LLM.
Why it matters
Unlike conventional programs that mainly judge eligibility, the generative AI also gives written reasons for exclusion, such as how many months until a patient meets a required observation period, which may ease the burden on doctors and trial coordinators.

WHO IT HITSDoctors and clinical-trial coordinators who currently check electronic medical records one patient at a time stand to benefit if the joint research bears out, as do pharmaceutical companies setting trial criteria.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Get these free every morning

Context & Analysis

NTT Data's existing screening service, launched in July, already receives trial criteria from pharmaceutical companies and analyzes electronic medical records including test values and doctors' notes on treatment courses and past history. When a match is found, it returns the patient ID to the medical institution so doctors can invite the patient to join. That service has already partnered with 11 medical institutions and received applications from three pharmaceutical companies.

The joint research adds generative AI to this base. The body notes that finding trial candidates has traditionally been done by doctors and trial coordinators checking electronic medical records one by one, comparing them against complex criteria that differ by trial, which weighs on medical staff. The new work also examines how much working time can be shortened, not just how well candidates are extracted.

NTT Data's target of 1 billion yen or more in fiscal 2029 sales combines the existing service with generative-AI-based offerings, suggesting the company sees the AI work as an extension of a business it is already selling, according to the announcement.

FAQ

What does the generative AI add compared with NTT Data's existing service?: The earlier service, launched in July, uses SQL, Python and other programming languages to extract candidates. The generative AI adds written reasons for exclusion and suggestions such as when a patient might become eligible.
How much did the earlier screening cut the number of patients doctors had to review?: In a technical validation with Kindai University Hospital on lung-cancer trials, about 3,000 patients' data was analyzed. The number of patients needing detailed review fell by 97.4% and 99.4% in two trials.
What is NTT Data's revenue target for this business?: NTT Data aims for sales of 1 billion yen or more in fiscal 2029, combining this service with generative-AI-based services.

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Get it free

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.