Before you ask AI anything about your cancer, read this first

A practical guide to using AI tools the right way — as a research companion, not a replacement for your oncologist.

Let me be direct with you: AI tools like ChatGPT, Claude, and Gemini are genuinely useful for cancer patients. They can help you understand complex medical language, organize your questions before an appointment, and make sense of research that would otherwise take days to decode. But — and this is important — most patients are using them wrong, and that gap can lead to confusion, false reassurance, or unnecessary panic.

As a researcher who studies how information moves between the medical system and patients, I’ve seen this firsthand. The problem isn’t that AI is unreliable. The problem is that AI, without the right context, is answering general questions when what you actually need are specific answers tied to your diagnosis, your cancer’s behavior, and your treatment protocol. Think of it this way: asking a general AI “what’s the best treatment for breast cancer?” without feeding it any of your actual clinical documents is like asking a brilliant doctor who has never seen your chart to make a recommendation on your treatment.

So before you open ChatGPT and start typing questions about your diagnosis, here’s what you actually need to do first.

Your oncologist’s opinion comes first. Always.

Before we get into checklists and tools, I want to anchor this entire conversation in one non-negotiable principle: no AI recommendation, no matter how confident it sounds, should come before the guidance of your oncology care team. AI does not know your full history, your bloodwork trends, how your body responded to prior treatment, or the dozens of judgment calls your care team has made on your behalf. An AI can explain what a treatment protocol involves. It cannot tell you whether that protocol is right for you.

Use AI as a tool to become a better-informed patient — so you can ask better questions, understand the answers, and participate actively in your own care. That’s the goal. Not to replace your team, but to walk into every appointment ready.

Why AI needs your documents, not just your questions

Here’s something that surprises most people: AI tools like ChatGPT and Claude don’t automatically know anything about your case. Every conversation starts fresh. When you type “I have stage III ovarian cancer, what should I know about my treatment?”, the AI pulls from general training data — not from established treatment algorithms, your specific pathology, or your institution’s clinical protocols.

To get responses that are actually meaningful, you need to feed the AI the right source material. Think of it as briefing a very capable research assistant who knows a lot about medicine in general but nothing about your case specifically. The better the briefing, the better the output.

This is where most patients go wrong. They skip the briefing and go straight to the questions. The checklist below is designed to help you build that briefing properly — before you type a single question.

Your pre-AI checklist: what to gather before you start

1. Your pathology report

This is the document that defines your cancer at the cellular level — tumor type, grade, staging, biomarkers, receptor status, and more. It’s often several pages long and dense with technical language. This is exactly why AI can help you parse it — but it needs to see the actual document to match it accordingly with the national treatment guidelines.

• Ask your office for a copy of this report

• The relevant treatment algorithm (PDF)

Treatment algorithms are the standardized, evidence-based decision frameworks oncologists use. The most widely referenced come from the NCCN (National Comprehensive Cancer Network). You can download NCCN guidelines for free by creating an account at nccn.org. Find the guideline that matches your cancer type and download it as a PDF. This document becomes the backbone of your AI conversations — it tells the AI what the standard-of-care options actually are for your specific cancer type and stage.

• These guidelines are what every oncologist in the US follows.

• There are two versions of the guidelines (one for physicians and one for patients).

• I recommend downloading the one for physicians and uploading that to your AI tool.

• Go to this website: https://www.nccn.org/guidelines/category_1

• Click on the first pdf "Guidelines" and download it - save this to your computer, as you may need it again

• Your treatment summary or care plan

Ask your oncology team for a written summary of your current or proposed treatment plan. This might include chemotherapy regimens, radiation dosing, surgical recommendations, or targeted therapy protocols. Having this in hand allows you to ask AI to cross-reference your actual plan against the treatment algorithm — so you can understand the reasoning behind what your team has recommended.

• If you have not discussed your treatment plan, this is still a great point to ask AI to help you understand your treatment options

• Recent lab work and imaging reports

Your CBC, metabolic panel, tumor markers, and imaging results (CT, MRI, PET scan reports) all provide context that sharpens the AI’s responses. You don’t need to share everything, but having the most recent reports on hand means you can paste in specific values or upload documents when asking targeted questions about your current status.

• I highly recommend using a note pad within your phone or a physical notebook that you bring with you to appointments to keep all the key information

• Check my resources page of this blog for a helpful note pad to copy/paste into your phone and start to fill out as you gain information on your diagnosis.

5. A list of your current medications and known allergies

Interactions between cancer treatments and other medications are clinically significant. Including your current medication list when asking AI questions about treatment options gives it the context to flag potential concerns — which you can then bring directly to your care team for confirmation.

How to actually use the AI: a step-by-step approach

Once you have your documents ready, here’s how to structure your AI sessions effectively.

1. Start with a clear briefing. Open a new conversation and tell the AI who you are and what you have. Something like: “I’m a patient with [cancer type and stage]. I’m going to upload my pathology report and the NCCN treatment guideline for my cancer type. Please help me understand the key terms in my pathology report and how my diagnosis maps to the treatment algorithm.” Then upload or paste your documents.

2. Ask specific, document-anchored questions. Instead of “what’s the best chemo for me?”, try “based on the NCCN guideline I uploaded and my pathology findings, what treatment categories apply to my staging? What are the first-line options listed?” The more specific the question, the more useful the answer.

3. Use AI to build your question list for your oncologist. At the end of every AI session, ask: “Based on everything we’ve discussed, what are the most important questions I should bring to my oncologist at my next appointment?” This turns your AI session into appointment preparation — which is exactly what it should be.

4. Never act on AI output alone. Every insight, concern, or option surfaced by AI goes back to your oncologist for validation. Write it down, bring it up, and let your care team respond. That loop — AI-assisted learning followed by expert human judgment — is where the real value lives.

Which AI tool should you use?

ChatGPT (from OpenAI), Claude (from Anthropic), and Gemini (from Google) are all capable of handling uploaded PDFs and having substantive conversations about complex medical documents. For most patients, the best tool is whichever one you find easiest to use consistently. Each has the ability to accept document uploads in their paid versions, which is important when you’re working with treatment algorithms and pathology reports.

That said, paid tiers generally offer better document handling, longer conversation context, and more consistent performance — all of which matter when you’re working through detailed clinical materials. If you’re going to rely on AI as a regular part of your cancer journey, the investment is worthwhile.

The bottom line

AI is a powerful tool for cancer patients who know how to use it. But like any tool, its value depends entirely on how it’s set up. Gather your documents first. Ground every conversation in your actual clinical materials. Let your oncologist make the calls. And use AI to show up to every appointment more informed, more prepared, and more empowered to advocate for yourself.

This post is for educational purposes only and does not constitute medical advice. Always consult your oncologist or qualified healthcare provider before making any decisions about your treatment.