When you are diagnosed with cancer, every instinct says do more. Bigger surgery, stronger chemo, the newest drug. It feels like fighting harder. For many cancers, doing more genuinely means living longer, and I would never talk anyone out of treatment that works.
But in a surprising number of well-run randomized trials, doing more did not help people live any longer. Sometimes it only added harm. These are trials have quietly reshaped how some oncologists think, and most of us have never heard of. The pattern touches on multiple kinds of treatments, so let’s get into it.
Surgery
Start with the most intuitive one, cutting the cancer out.
For most of the twentieth century, surgeons believed breast cancer spread outward in an orderly march, so more radical surgery should mean more cures. In the NSABP B-04 trial, women were randomized to radical mastectomy (which removes the breast along with the underlying chest wall muscles and all the axillary nodes) or to a total mastectomy (which leaves the muscles intact and skips the routine node dissection), and the survival curves were identical.
The twenty-five-year results, published in 2002, showed the same thing. The conclusion was that breast cancer is often systemic from early on. Cells have frequently already left the breast by the time we find the tumor, so how aggressively you attack the local site does not change the outcome. If B-04 dismantled the radical mastectomy, the B-06 trial then proved that lumpectomy was a safe option.
The same lesson came for the axilla. For decades, a positive sentinel node meant a full axillary dissection, which carries a real risk of lifelong lymphedema. Published in 2017, the ACOSOG Z0011 trial randomized women with one or two positive sentinel nodes to full dissection or none. At ten years, overall survival was the same!! 86 versus 84 %. Removing the extra nodes bought arm swelling, not time. And many women with limited nodal disease keep their axilla today.
Next came the metastatic kidney cancer discovery. In the 2018 CARMENA trial, removing the primary tumor in patients whose cancer had already spread did not extend survival over drug therapy alone.
And, in melanoma, the 2017 MSLT-II trial found that removing all remaining lymph nodes after a positive sentinel node did not improve survival. It produced more lymphedema, not more life.
Chemotherapy
For decades, node-negative, hormone-positive breast cancer meant chemotherapy for almost everyone, because we didn’t know who benefited, so everyone got the “just in case” chemo. Genomic testing changed that. In the 2018 TAILORx trial, women were assigned chemo or no chemo based on a 21-gene recurrence score. For the large midrange group, adding chemotherapy did not improve survival. The 2016 MINDACT trial reached a parallel conclusion using a different genomic signature. today that means that a great many women were given chemotherapy, with all its toxicity and all its ugliness, that never changed their odds of living longer.
The deeper point is that adjuvant chemotherapy, chemo given after surgery to lower recurrence, is measured in relative risk reduction, which can sound impressive while the absolute benefit is small. A 1/3rd drop in recurrence sounds like a lot, but for a low-risk patient, say 5 to 15 %, it can mean a single percentage point of actual survival gain, sometimes less. For higher-risk disease the absolute benefit is real and larger. Who (you?) decides if that’s worth it?
Probably, the most dramatic example here is from the 1990s. Tens of thousands of women with breast cancer received high-dose chemotherapy with an autologous bone marrow or stem cell transplant, an ordeal that killed a meaningful fraction of them outright. The theory was that more chemo, pushed to the edge of what the body could survive, would wipe out more cancer. Not so. High-dose therapy did not improve survival over standard chemotherapy. Median survival was 24 months versus 26.
I recently read a lovely piece by Cancer Ecology by Ken Pienta, that names the mechanism underneath a lot of this. A tumor is not one uniform mass. It is billions of genetically different cells under constant selection. When we hit it with a drug, we are not teaching it to resist. We are clearing the field for the rare cells that were already resistant, then letting them repopulate. Push the same pressure harder and longer and you often make it worse, because you are running an evolution experiment and evolution is very good at those. Read full article below -
Genetic testing that can spare you harm
There is a different kind of “more” worth questioning. Some people carry inherited enzyme variants that make standard chemo doses toxic, sometimes fatally so. A cheap blood or cheek swab test before treatment can identify them, yet preemptive testing is still rarely done.
One example is DPYD. This gene makes the enzyme that clears fluoropyrimidine chemotherapy, the 5-FU and capecitabine backbone used in colorectal, breast, and many gastrointestinal cancers. People with a DPYD variant break the drug down slowly, so a standard dose behaves like an overdose and can cause severe, occasionally lethal toxicity. Testing first lets us reduce the dose to a safe level. Europe now recommends DPYD testing before fluoropyrimidines. The United States has been slower, per usual, and only far and few clinics are testing this on the regular.
Others include UGT1A1 testing before irinotecan (most often used in colorectal and pancreatic cancer); mutation in UGT1A1flags people at risk for severe diarrhea and neutropenia.
TPMT and NUDT15 testing before thiopurines like 6-mercaptopurine and thioguanine, used in leukemia, can explain up to half of serious toxicities and is standard of practice.
G6PD testing before rasburicase and high dose vitamin C IVs prevents dangerous hemolysis.
CYP2D6 status affects how well tamoxifen is activated, though the clinical data there is somewhat mixed
CYP2D6 also drives ondansetron and opioid metabolism, so poor and ultrarapid metabolizers get unpredictable antiemetic and pain control.
CYP2C19 affects voriconazole antifungal dosing, which comes up constantly in neutropenic patients.
One panel can cover all of these. CPIC-guided pharmacogenomic panels run DPYD, UGT1A1, TPMT, NUDT15, CYP2D6, and often G6PD off a single sample.
CPIC publishes the dosing guidelines for every one of these and is the best reference I have found.
Getting pharmacogenomic testing is not about doing more. It is about not being harmed by a dose that was never right for your biology.
Testing the tumor, not just you
There is a second kind of genomic testing, and it looks at the cancer rather than your inherited genes. Sequencing the tumor, either from a tissue biopsy or from a liquid biopsy that picks up circulating tumor DNA in a simple blood draw, can find specific alterations that predict whether a targeted drug will actually work. It cuts both ways. A match can open a treatment built for your biology. No match can spare you a drug that was never going to help. Liquid biopsy is easy to repeat over time and can flag resistance as it emerges, though a tissue sample is still more sensitive for the first full profile. Before starting any targeted therapy, ask whether your tumor has been profiled.
Drugs approved without proof you will live longer
Ahhh, a topic quite possibly the most mind boggling of them all. Many cancer drugs are approved by regulators on surrogate endpoints, meaning the tumor shrank or growth was delayed on a scan, not proof that patients lived longer or better. Progression-free survival and response rate are easier and faster to measure than survival, so that is what many approvals are based on.
The problem is that a shrinking tumor does not reliably translate into a longer (or happier) life. When researchers followed cancer drugs approved on these surrogates through their confirmatory trials, only about one in five went on to actually demonstrate that patients lived longer.
Bevacizumab (most commonly used to treat colorectal cancer, lung cancer, and ovarian cancer) was granted a metastatic breast cancer indication on delayed progression, then the indication revoked in 2011 when the survival benefit failed to materialize and the toxicity was real. The drug looked good on a scan but not so much in real life.
In 2021 the FDA ran an industry-wide review of these “conditional” approvals, and many of them were retracted.
Atezolizumab lost both its metastatic bladder cancer and its PD-L1-positive triple-negative breast cancer indications when the confirmatory trials failed to show a survival gain.
Durvalumab was pulled in bladder cancer for the same reason.
Nivolumab was withdrawn in third-line small-cell lung cancer and in previously treated liver cancer.
Pembrolizumab in third-line stomach cancer.
Melphalan flufenamide, arguably most alarming, was approved in multiple myeloma in early 2021 on progression-free survival, its confirmatory trial showed patients on the drug actually died sooner, 19.7 months versus 25.0. The drug was off the market within months.
This does not mean these drugs don’t help. Some control symptoms, delay progression, and make a meaningful difference in quality and duration of life. Several of these same molecules are lifesaving in other cancers. It means the label “approved” does not automatically mean “helps you live longer,” and it is fair to ask which one you are getting.
Treatment at the very end of life
The hardest example is chemotherapy in the last weeks of life. It is often given in the hope of buying time. Does it?
In a 2015 study of patients with advanced cancer and good performance status, chemotherapy near the end did not extend survival, and it worsened quality of life in the final weeks.
Radiation shows the same pattern. For painful bone metastases, multiple randomized trials found that a single dose of radiation relieves pain as well as ten, yet the longer course is still routinely given.
Meanwhile, a few trials suggest that adding palliative care early, alongside treatment rather than only at the end, improves quality of life and in some cases survival. Patients who got it felt better, chose less aggressive care near death, and sometimes lived longer.
What about screening and treating slow cancers
One more version of “more” deserves a mention, because it starts before treatment. Some cancers grow so slowly that finding and treating them changes nothing except how much harm you absorb. The ProtecT trial followed 1,643men with localized prostate cancer randomized to active monitoring, surgery, or radiation. At fifteen years, prostate cancer survival was about 97 percent in every group. Surgery and radiation cut the risk of the cancer spreading, but that did not translate into living longer, and both carried real costs to urinary and sexual function. For many low-risk prostate cancers, watchful monitoring is now the standard, not a compromise.
The same logic increasingly applies elsewhere. Small, low-risk thyroid cancers are now often watched rather than removed. In chronic lymphocytic leukemia, watch and wait has been standard for decades, since treating early adds toxicity without extending life. Low-tumor-burden follicular lymphoma follows the same rule, with immediate treatment showing no survival gain over monitoring. Small kidney tumors under about four centimeters are increasingly followed with serial imaging rather than surgery, especially in older patients. Smoldering myeloma and MGUS are watched rather than treated, because most never progress. And some cases of DCIS in the breast are now being tested against active monitoring (trials like LORIS and COMET).
I believe this is where integrative care, diet and lifestyle can make the biggest impact. When the standard is to monitor rather than cut or irradiate, that is a window for real meaningful change and the best chance we get to alter the terrain the cancer sits in. Indolent tumors are shaped by their environment, and much of that environment is modifiable. Metabolic function, insulin resistance and glucose control, chronic inflammation, immune surveillance, and hormone balance all influence whether a slow cancer stays slow.
What you can do
Before any major treatment, ask the direct question: does this improve my survival, my symptoms, or neither, based on trial data for my exact cancer and stage. All three can be valid reasons to treat. They are not the same reason.
For early breast cancer, ask whether a lumpectomy or a smaller axillary surgery is an option before agreeing to a bigger operation. Removing more tissue has not meant more life.
For hormone-positive, node-negative breast cancer, ask whether Oncotype or MammaPrint can tell you if chemotherapy would actually help before you agree to it.
For any adjuvant chemo, ask for the absolute benefit, not the relative one. A one-third drop in recurrence can still mean a single percentage point of survival. Then decide whether that is worth it to you.
Before fluoropyrimidines, irinotecan, thiopurines, rasburicase, or high-dose IV vitamin C, ask whether you have been tested for the relevant enzyme variant: DPYD, UGT1A1, TPMT, NUDT15, or G6PD. One CPIC-guided panel covers them.
Before any targeted therapy, ask whether your tumor has been profiled, by tissue or liquid biopsy, to confirm the drug matches your biology or if there’s a better match available
For any new or expensive drug, ask whether it has been shown to help people live longer, or only to shrink the tumor or delay progression on a scan.
If you are facing advanced disease, ask about early palliative care alongside treatment, not only at the end.
For painful bone metastases, ask whether a single dose of palliative radiation will do.
For a slow-growing, low-risk cancer in active monitoring get an integrative provider on board and use that window for metabolic, inflammatory, immune, and hormonal work.
Get a second opinion for any non-emergent cancer treatments
Ksenia Malarkey, ND is a naturopathic physician specializing in integrative oncology, hormones, and metabolic health.
She provides medical care for Washington residents and offers global health coaching.
To schedule a free 15-minute consult, visit drkseniamalarkey.com