The client: a US cancer-care company
Our client is a US cancer-care company whose software is used by oncology departments in hospitals and clinics across the United States. Through those partnerships it holds millions of historical clinical records: diagnoses, cancer stages, procedures, medicines and how each patient's treatment went. Its care pathway also includes a separate patient app, where people in treatment report symptoms, answer questionnaires from their care team, keep track of appointments and results, and message nurses between visits.
Why millions of oncology records were hard to use
That history could help treat the next patient, but it was hard to use. Records arrived from many hospitals and clinics in different formats and with gaps, so nobody could easily ask the question that matters most: which past patients looked like this one, at this stage, and what happened on each treatment path they took? Day to day, oncology teams also lacked a single view of their own work: missed appointments, open cases, treatment plans waiting for approval, and whether the data behind it all could be trusted.
What we built: data lake, ML models and oncology dashboards
Clinical data lake and pipelines
Pipelines that pull clinical records from partner hospitals and clinics, clean and standardise them, score their completeness, and load them into one data lake ready for analysis.
Similar-patient analysis
Python machine-learning models that find past patients with a similar profile, diagnosis, stage and treatment history to a new patient, so clinicians can see which treatment paths were tried and how each one turned out.
Outcome and risk models
Models trained on historical treatment data that estimate how different treatment paths are likely to go, giving oncology experts evidence to weigh before they choose one.
Oncology dashboards in Power BI and Tableau
Role-based dashboards for each team's day: activity and missed appointments, case management against targets, treatment plans awaiting approval, treatment volumes and a data quality scorecard for every site.
How TechTose worked with the client's oncology experts
We worked alongside the client's oncology experts, clinicians and stakeholders, led by our CTO. The clinicians defined the questions that matter for treatment, and TechTose turned them into pipelines, models and dashboards, then reviewed the results with them. Because we owned every layer, from the data lake to the models to the dashboards, a new clinical question could go from idea to a working view without a hand-off between teams. Records, code and client details stayed confidential under the NDA throughout.
How machine learning supports treatment decisions
Machine learning sits behind the clinical side of the platform. When a new patient is assessed, the models surface similar past cases and how each treatment path went, and the outcome and risk models add an estimate for the options under consideration. The decision always stays with the oncology team: the models give them evidence from millions of past records, not an answer.
Where the platform stands now
Our work on the platform was completed in 2024, and the data lake, models and dashboards are in the hands of the client's own team.
Results
- Millions of historical clinical records from US hospitals and clinics in one data lake, ready to analyse
- Clinicians can compare a new patient with similar past cases and see how each treatment path went
- Day-to-day oncology operations in one set of dashboards, with data quality scored for every site
Why the client and product aren't named
Many companies hire us to build the product their business depends on. Before we see a line of code, they ask us to keep it private, and we agree. So this page describes the work and the problem it solved, but not who it was for.
How we handle confidential work
- We sign your NDA before you share code, data or plans.
- Your code and accounts are yours. We work inside them, and our access ends when the work does.
- We never name a client, product or person publicly without written permission.
- We describe the work, not the client, as we have on this page.
Want the details? On a call we can walk you through this project, within what the NDA allows.
Questions about this oncology analytics project
Why isn't the client named?
We signed an NDA, so the client's name, its products and its code stay confidential. We protect every client's work the same way. If you'd like to know more, book a call and we'll share what the NDA allows.
How can machine learning help treat future cancer patients?
By learning from the patients treated before. In this project, TechTose built models that find past patients similar to a new one, by diagnosis, stage and treatment history, and estimate how different treatment paths are likely to go. Oncology teams use that evidence alongside their own judgement, and the decision stays with them.
What did TechTose build for this oncology analytics platform?
A clinical data lake fed by pipelines from US hospitals and clinics, similar-patient and outcome models in Python, and Power BI and Tableau dashboards for activity, case management, treatment planning and data quality.
Does TechTose build healthcare analytics and machine-learning software?
Yes. This project covered the full path from raw clinical records to models and dashboards that oncology teams use. Most projects start with a free Scan of your app, code or idea, which comes back as a written report in 48 hours.