Independent ecancer blogger Dr Bishal Gyawali offers his reflections on the results of the MINDACT study.

There is a fundamental limit to the precision with which both the complimentary variables of a particle can be known. There is a minimum for the product of the uncertainties of these two measurements such that precision in measuring one variable increases the uncertainties in the other.
-Heisenberg’s Uncertainty Principle

The MINDACT study was a fantastic collaborative act of researchers from 112 institutions in 9 European countries, enrolling 6693 women with early breast cancer and planning to follow them up for at least 10 years (15 for those on hormone therapy). Truly an amazing feat and big congratulations! The 5 year results from the MINDACT study have now been published in the New England Journal of Medicine accompanied by an editorial and a perspective.

The decision to pursue a course adjuvant therapy in patients with early breast cancer is not straightforward. The standard practice has been to prescribe anti-HER2 therapy plus chemotherapy in HER2-positive cancers, chemotherapy alone in triple negative cancers, and endocrine therapy - with or without chemotherapy - in hormone-positive, HER2-negative cancers.

It is in this hormone positive, HER2 negative subgroup of cancers that the biggest confusion lies. Many patients do well with endocrine therapy alone and thus, the addition of chemo could be overtreatment: lack of clinical benefit despite the toxicities of chemotherapy. On the other hand, forgoing chemo could be under treatment in a subgroup of patients who are at higher risk of relapse and poor survival.

Hence, experts and guideline groups have suggested some possible “high-risk” clinical features such as larger tumour size, higher proliferation index markers (Ki-67), involvement of lymph nodes and young age to indicate aggressive disease that could possibly derive the most benefit with the addition of chemotherapy. Indeed, Internet tools such as Adjuvant! Online have been developed incorporating these and other factors to determine whether a patient falls into the “higher-risk” or the “lower-risk” group for relapse to help make decision of adding/forgoing chemotherapy.

As we began to understand the biology and the genomics of breast cancer in more detail, quantifying recurrence or survival probabilities became easier with the introduction of technologies such as Oncotype Dx and Mammaprint that interrogate the genomic signatures in an individual tumor and provide an assessment of the relative risk of progression or survival. These genomic tests are now recommended by the guidelines to assess the prognostic risks and thereby help decide if the benefits of adding chemotherapy would be worth the toxicities, especially in hormone positive HER 2 negative tumours. However, the predictive utility of these tests in deciding appropriate therapy has never been prospectively studied.

Last year, we saw the results from a study that prospectively studied the Oncotype Dx test in women with early hormone-positive, HER2 negative, node- negative breast cancer who would normally have been offered adjuvant chemo by the NCCN guidelines.   These women (n = 1626) underwent the Oncotype Dx test, and those with a low score of 0 to 10 were provided hormone therapy alone without chemo. At 5 years, they had an invasive disease free survival rate of 93.8%, metastasis free survival rate of 99.3%, recurrence free survival rate of 98.7% and an overall survival rate of 98.0%. Although this was not a randomised study, a less than 1% risk of distant metastases and 1.3% risk of recurrence gave confidence to many oncologists to skip chemotherapy in these patients.

However, these patients don’t represent the majority of the patients, where genuine clinical doubts still exist. Hormone-positive, HER2-negative, node-negative breast cancer represents a very favourable subgroup and hence may show such better survival results. Whether chemotherapy can be safely omitted in high-risk patients with involved lymph nodes or positive HER2 status is a more difficult question to answer.

The MINDACT study attempted to answer these difficult questions. The investigators tested if there were some true detrimental effects when chemotherapy was skipped in clinically high risk patients based on the genomic risk categorisation. They also included even HER2-positive and triple-negative patients in this study - a group for whom one would usually prescribe a straightforward course of chemotherapy. Indeed, the NCCN guidelines and the St. Gallen consensus guidelines recommend chemotherapy for HER2-positive and triple-negative subsets. Thus, the investigators must be applauded for daring to include these patients as well in their trial, and to test whether chemo could be safely omitted for those with low genomic risks.

Also noteworthy: unlike the Oncotype Dx trial, the MINDACT was a randomised trial.

In the MINDACT trial, the investigators tested the clinical risk (C-R) of patients using the standard Adjuvant! Online tool and genomic risk (G-R) using the Mammaprint. For concordant results, the decision was straightforward: C-R low, G-R low received no chemo; C-R high, G-R high received chemo.

But when the results were discordant (C-R high, G-R low or C-R low, G-R high), the patients were randomized to chemo or no chemo arms. The primary objective was to show the non-inferiority of forgoing chemo in the C-R high, G-R low group and this would be achieved if the lower limit of the 95% CI for distant metastasis free survival rate (DMFS)was 92% or higher. It would imply that G-R is more important than C-R in making treatment decisions for this subgroup. The study met its primary endpoint. The DMFS for patients who didn’t receive chemo was 94.7% (95%CI 92.5% to 96.2%) at 5 years. In the analysis of intention-to-treat , the patients who received chemo versus those who didn’t had a better 5 year DMFS rate by a non-significant 1.5 % ( 95.9% v 94.4%, HR 0.78, p = 0.27). Another equally important finding is that for C-R low but G-R high patients, there was no advantage to adding chemotherapy with a 5Y DMFS rate of 95.8% in patients who received chemo versus 95.0% in those who didn’t ( HR 1.17, p = 0.66).

No differences in OS were noted in either group. The only significant finding as seen from Table 2 is in the disease-free survival rates: chemo group fares significantly better than no chemo group (93.3% v 90.3%, HR 0.64, p = 0.03). This study also validated the prognostic importance of the test as the test was significantly associated with DMFS in a multivariable analysis after adjusting for other variables.

One commendable job the investigators have done in this trial is the use of Adjuvant! Online for categorising C-R. This minimises physician’s subjective bias in risk categorization and improves the objectivity and reproducibility of the study. The authors describe the study in good detail with ample data in the supplementary appendix and highlight the inadvertent error in G-R calculation for a certain time period. That error was accounted for in the analysis and sample size was re-adjusted. Secondly, the authors present the data in a relatively easy-to-understand way. Of the 3356 pts who were C-R high, 1550 (46.2%) had low G-R. Thus, with G-R assessment, nearly 46% of patients would be spared the toxic effects of chemo without any detrimental impact.  However, readers like myself would have loved more infographics and data in terms of number needed to treat (or harm) for easy communication with patients.

So, are these results practice changing? Do they apply to my patient?

As is the case with all non-inferiority studies, the cut-off limit for non-inferiority margin is arbitrary. Is a 5-year DMFS rate of 92% enough to decide that chemo can be safely excluded? This can be argued, and the opinion will differ between individuals. In addition, a nonsignificant but 1.5% better DMFS rate at 5 years was observed in the chemo group, when compared with no chemo group. However, before making interpretations from the secondary objectives, it is essential to note that this study was not powered to make statistical comparisons between the randomised groups. 

The MINDACT study provides an excellent example of the complexities surrounding precision medicine.  First, it shows us the limitations of precision medicine. Precision medicine or genomics cannot replace the clinic-pathological assessment. Precision medicine can only be an addition not a substitute, only a complement not the whole of clinical management. For C-R low, G-R high patients, there was no benefit with chemo. Thus, GR alone should never guide the treatment plan.

Second, more precision brings more uncertainties. With medicine getting more precise, patient care is getting more and more imprecise. When a patient arrives in clinic tomorrow to discuss her options of adjuvant treatment, we may have precise knowledge on the genomics of her cancer but only uncertainties in recommending treatment options. As the perspective that accompanied the MINDACT study pointed out, many healthcare professionals are ill-equipped to communicate this uncertainty.

How do we get the data from this study to help the patient make decisions? We considered 92% to be non-inferior limit, but would the patient? We might think 1.5% difference is not statistically significant and the patient can safely avoid the toxicities of chemo; but would she consider a probable 1.5% difference in 5 years important enough to undergo chemo? These are always going to be subjective decisions. Adding to the uncertainty, would providing chemo even for low risk patients falsely reassure her and decrease the compliance with hormone therapy?

Thus, I suggest that this genomic test should be a part of informed/shared decision making. While patients at clearly low C-R might omit both genomic testing and chemotherapy, those at high C-R might find this genomic test useful. For patients with high C-R ready to undergo chemotherapy anyway, further genomic testing wouldn’t be necessary as it wouldn’t change the treatment plan. This test will have the greatest benefit for those who would like to avoid chemotherapy but would want a confirmation that their future risk won’t rise significantly by doing so. Detecting low G-R based on the test would allow them to avoid chemotherapy with confidence. An important caveat, however, is that the genomic risk in this study is simply dichotomised into low and high while in reality it is a continuum.

I remember Heisenberg’s Uncertainty Principle from my high school Physics lecture. Heisenberg (Nobel Prize in Physics, 1932), in his famous Uncertainty Principle, stated that the position and momentum of a particle cannot be measured simultaneously with precision. Precision in the measurement of one reduces the precision in the measurement of the other. I find his principle to be very true in cancer medicine.

In the old days when there was no “precision medicine” on the horizon (imprecision in cancer diagnostics and biology), we prescribed chemotherapy to everyone without any dilemma (confidence in treatment). As we have now improved our understanding of cancer and achieved better precision in categorising cancers based on genetic signatures, we are becoming unsure of whether a particular treatment would be effective for the patient, as we now know that the genomic landscape of individual tumour in an individual patient is unique.

Only with more knowledge comes the realisation of one’s limitations. True wisdom is the humility to accept the uncertainties and still strive to help the patient make an informed choice in trading off the risks and benefits with each option. The true goal of precision medicine would not be to provide a definite answer to every question, but to broaden the uncertainties and provide more probabilities for the patient and physician to work together with to take the best bet of a treatment option that caters to the patient’s needs.

Bishal Gyawali (MD) is an independent blogger. He is undergoing his postgraduate training in medical oncology at the Graduate School of Medicine, Nagoya University, Japan, where he is also a PhD candidate under the Japanese government scholarship. He also serves as visiting faculty at the department of Hemato-Oncology in Nobel Hospital, Kathamandu, Nepal. He graduated in medicine from Institute of Medicine, Tribhuwan University, Nepal in 2011 with seven gold medals for his academic excellence. He has been honoured with “Student of the Decade award” and “Best Student Award” for his academic excellence in Nepal. His areas of interest include evidence-based oncology practice, cost-effectiveness of cancer therapies and economic feasibility of cancer management in low-income countries. Dr Gyawali tweets at @oncology_bg. Dr Gyawali is an independent blogger and his views are not representative of ecancer.

Read Dr. Gyawali's previous blogs here.