Questions to Ask About Breast Cancer Treatment

Before any treatment question makes sense, three test results have to be on the table. NCI lists them as the core of molecular profiling in breast cancer. They are estrogen receptor (ER) status, progesterone receptor (PR) status, and HER2 receptor status. HER2 stands for human epidermal growth factor receptor 2.

Those three results sort breast cancer into one of three types:

• Hormone receptor positive.
• HER2 positive.
• Triple negative — ER, PR, and HER2 all negative.

NCI states plainly why this matters. ER, PR, and HER2 status help set the prognosis. They also predict response to endocrine therapy and HER2-directed therapy. A drug that works on one type does nothing for another.

There is also a quality-control question worth asking. The American Society of Clinical Oncology and the College of American Pathologists have published consensus guidelines here. Those standardize how these assays are performed, interpreted, and reported. ER and PR are measured by immunohistochemistry. HER2 is measured by immunohistochemistry and in situ hybridization. So it is fair to ask which method produced a result, and whether it was borderline.

Beyond the receptors, gene profile tests estimate recurrence risk and help decide about chemotherapy. NCI names three: MammaPrint, Oncotype DX, and the Breast Cancer Index.

MammaPrint was the first gene profile test approved by the FDA. It is a 70-gene signature. It sorts tumors into high-risk and low-risk categories. The MINDACT trial, registered as NCT00433589, set out to test its usefulness. The question was patient benefit for adjuvant chemotherapy decisions.

Oncotype DX has the most extensive clinical validation so far. It applies to HER2-negative, hormone receptor-positive breast cancer. It produces a 21-gene recurrence score. That score comes from the expression levels of those genes.

Those score thresholds are worth knowing exactly, because they are decision points rather than descriptions. For node-negative disease:

• Recurrence score of 11 or less: low risk. Chemotherapy is not indicated.
• Score above 11 and up to 25: intermediate risk. NCI calls the decision here complex and personalized. Age at or under 50 versus over 50 feeds into it. So do clinicopathological features and patient preference.
• Score above 25: high risk. Chemotherapy is indicated.

For postmenopausal node-positive disease the cutoff is simpler. A score of 25 or less is low risk, and chemotherapy is not indicated. A score above 25 is high risk, and chemotherapy is indicated.

The Breast Cancer Index combines two profiles. One is the HOXB13 to IL17BR expression ratio, called the H over I ratio. The other is the Molecular Grade Index. It has been both prognostic and predictive in hormone receptor-positive disease.

One older dataset shows why these scores exist. In the NSABP B-14 trial, tissue was available from 668 patients treated with tamoxifen. The 10-year distant recurrence risk was 7% with a low recurrence score. It was 14% with an intermediate score, and 31% with a high score, at P less than .001. That trial used different thresholds from current practice. They were under 18, 18 to 30, and 31 or above. That gap is itself a reason to ask which cutoffs a specific report uses.

This section tends to surprise people, and it comes straight from randomized trials.

NCI calls one thing controversial. That is the frequency of follow-up, and the value of screening tests, after primary treatment for stage I, II, or III breast cancer. Then it gives the finding. Randomized trials show that periodic bone scans, liver ultrasound, chest X-rays, and liver function blood tests do not improve survival or quality of life. The comparison is with routine physical examinations.

And it goes one step further. Even when those tests permit earlier detection of recurrent disease, patient survival is unaffected.

On that basis, NCI names what acceptable follow-up can be limited to. This applies to people with no symptoms who have finished treatment for stage I to III disease. The list is two items. Physical examination, and annual mammography.

That is a short list. It is worth discussing openly. A plan with fewer scans is often mistaken for less care.

Locoregional recurrence means the cancer returns in the breast, chest wall, or nearby nodes rather than at a distant site.

The rates have fallen over time. A meta-analysis suggests a rate under 3% after breast-conserving surgery plus radiation therapy. Rates are somewhat higher after mastectomy, up to 10%.

Two facts shape what happens next. Between 9% and 25% of people with a locoregional recurrence already have distant spread, or locally extensive disease, when it is found. That is why restaging to check the extent of disease comes before treatment. And confirmation by cytology or histology is obtained whenever possible, rather than treating on imaging alone.

Then comes the detail that changes drug choices. ER, PR, and HER2 status at the time of recurrence are weighed alongside past treatment. NCI notes that ER status may change at the time of recurrence. A receptor profile from years ago may no longer be the one that counts.

NCI adds two things about outlook. Recurrent breast cancer often responds to therapy, though treatment is rarely curative at that stage. And people with locoregional recurrence may become long-term survivors with the right therapy.

For 2026 in the United States, the American Cancer Society projects 321,910 new cases of invasive breast cancer in women. It projects 42,140 deaths, plus 60,730 cases of ductal carcinoma in situ. These annual counts come from Cancer Facts & Figures, and NCI's PDQ summary reprints them, currently from the 2025 edition. Fewer than one in eight women diagnosed with breast cancer will die of it. For comparison, about 61,950 American women will die of lung cancer in 2026.

Men account for 1% of breast cancer cases and 1% of breast cancer deaths.

On inherited risk, 5% to 10% of all women with breast cancer may carry a germline pathogenic variant in BRCA1 or BRCA2. Two risk-estimation tools are in common use. One is the Gail model. The other is the IBIS/Tyrer-Cuzick model, version 8. That version weighs family history more fully than the Gail model, and also accounts for breast density.

• What are the ER, PR, and HER2 results, and by which method were they measured?
• Was any of those results borderline or equivocal?
• Which type is this — hormone receptor positive, HER2 positive, or triple negative?
• Is a gene profile test being ordered, and which one?
• If Oncotype DX: what is the recurrence score, and which cutoffs apply given node status and menopausal status?
• For an intermediate score, what specifically tips the chemotherapy decision here?
• What will follow-up consist of, given that added scans and blood tests have not improved survival?
• Has germline genetic testing been discussed?
• At recurrence: has the receptor status been retested on the new tissue?

For how stage maps onto treatment, see breast cancer treatment by stage and breast cancer stages. This page is a question list drawn from NCI. It does not recommend a treatment.

• NCI PDQ — Breast Cancer Treatment, health professional version
• NCI — Breast Cancer Treatment by Stage
• American Cancer Society — Cancer Facts & Figures