C the Signs explained how AI helps detect cancer at the earliest stages

Bea Bakshi, head and co-founder of C the Signs, spoke about the future of oncological diagnostics, emphasizing not treatment, but the moment when the disease can still be caught earlier. According to her, AI is already becoming a working tool for early warning: it does not make a diagnosis on its own, but helps to notice risk faster and not lose the patient at the first stage.

Why the start matters

In oncology, timing often decides more than the loudness of symptoms. At early stages, many types of cancer disguise themselves as ordinary complaints: fatigue, weight loss, appetite changes, recurring pain that is easy to blame on stress or other conditions. The problem is not only that a single symptom seems insignificant, but that the doctor and the system need to timely put together many weak signals into one picture.

This is exactly where digital tools can provide a noticeable advantage. Bakshi essentially describes AI as an additional layer of attentiveness within primary care. Instead of waiting for the clinical picture to become obvious, algorithms can suggest earlier that a set of complaints, medical history, and risk factors require checking.

This approach is especially important where doctors have little time for appointments, and the patient's path to specialized examination already stretches across weeks, months, and several follow-up visits.

Where AI helps

This is not about a "magic button" that finds cancer without human involvement. The strength of AI is different: it can quickly compare scattered data and highlight cases that deserve increased attention. In early detection, this can work on several levels at once — from the first complaint in a therapist's office to the decision of whether a patient needs an expedited route to tests, imaging, or consultation with a specialist.

• Analyze combinations of symptoms that individually look like everyday complaints
• Compare complaints with age, family history, and other risk factors
• Suggest when a patient should be referred more quickly for additional examination
• Reduce the probability that an alarming signal will be lost in the stream of routine appointments

For platforms like C the Signs, the value here is not in a beautiful demonstration of the model, but in a concrete clinical pathway. If the system helps the doctor assign the necessary test earlier or not miss a follow-up visit with the same complaints, this is already a practical effect. The earlier a justified suspicion arises, the greater the chances that treatment will begin before a severe stage, when the choice is no longer so wide.

Not instead of a doctor

At the same time, Bakshi does not present AI as a replacement for clinical decision-making. Even an accurate model works only in context: examination, medical history, test results, and understanding what is happening with a specific person, not with an average patient from the dataset, are important. Therefore, such systems have two main checks — usefulness in a real appointment and quality of routing, not just a high metric on a test sample.

AI is part of the solution already at the earliest stages.

There are also obvious limitations. Any early screening system must balance between sensitivity and the number of false alarms: if there are too many signals, the doctor stops trusting them; if the threshold is too high, an important case can be missed again. Therefore, the next stage for such products is not only to improve the models, but also to embed them in the work of clinics so that the suggestions are explainable, timely, and really accelerate the patient's path to diagnosis.

What it means

The main idea is simple: a breakthrough in oncology is not only new drugs, but also an earlier moment of intervention. If AI helps to see the risk before the disease became obvious, it brings value not at the level of hype, but at the level of missed and timely found cases.