Your homework:  compare and contrast three new medical editorials, about research in diagnostic imaging in Alzheimer Disease (AD), or diagnostic genomic testing in Breast Cancer, and if possible submit a bigger picture perspective of how AD patients can ever get to real therapy.  First some US stats: Breast cancer new cases ~ 252,000 and deaths ~40,600; Alzheimer’s new cases ~ 487,000 and deaths ~121,000.

Whaaat?  How can you compare genomic tests to radiology tests?  Both have positive and negative aspects. Two of the radiology tests are for molecular factors, but the real issue is about how much of a diagnostic description is necessary to try to treat a condition, and do we expect too much from these tests?  What level of information is necessary for action?

“Has the Genome Granted Our Wish Yet?”  in NEJM is about general genomic testing, using the example of predicting breast cancer risk, but also mentioning other adult polygenic conditions (heart disease, obesity, diabetes).  They use a card playing metaphor: “To a great extent, it is not only the genetic hand you are dealt but also how you play that hand that makes the big difference in health outcomes.” [hence the picture above, trying to predict the draw, but I’m not recommending gambling!]

“Putting the New Alzheimer…(AT[N]) Diagnostic System to the Test” in JAMA is a commentary that discusses a study using current research dementia radiology and correlating the findings to memory testing, to see how well three kinds of brain scans can predict problems. The authors of the study find “a small….improvement in predicting memory decline”, but also write “the clinical importance of the difference is uncertain,”  over and above old-school clinical neuropsychological evaluations. The editorial says “…the added value of each specific AT(N) measure has not been fully delineated, and requires evaluation in future work, as does whether the findings of the current study will generalize to other biomarkers (CSF, FDG) and more diverse patient populations.”

“The Precision of Evidence Needed to Practice “Precision Medicine”  in NEJM is a commentary about the study of a specific genomic prognostic test in a subset of patients who already have the diagnosis of breast cancer, HR+, HER2 neg, used to predict recurrence, finding that old-school tumor size, age and histologic grade can still be useful in making decisions about therapy, over and above this and other genomic tests. They say that “clinical acumen will continue to be necessary.”

The first editorial, by Hunter and Drazen, basically defines “clinical utility,” which “depends on the existence of interventions that reduce risk among people identified as high-risk,” and calls out genomic information that is “prognosis without promise,” which one assumes may just cause anxiety, long an ethical issue in genetic testing.  They point out that the first genomic tests were available over ten years ago, but suggest the tests are not the genii in the bottle that grants actionable wishes.

These five papers come from the work of diligent, brilliant researchers.  But the studies themselves almost seem like they’re from different decades.  The glaring difference, to me, is not about how “advanced” and nuanced cancer research questions and results are, with a impact that may save even more lives, but how AD research questions and answers seem mired in factors of uncertain significance .  Of course, the AD work is about vetting a diagnostic system, and not about therapy or outcomes at all; one need not mention AD single agent trial failures.

What is telling is that the breast cancer recurrence risk test discussed here is a 21-gene panel, 16 “cancer-related” genes and 5 reference genes.  That would seem to be precise.  Yet the results are reported as aggregate recurrence risk scores, implying that no one gene tells the whole story.  Those scores are even grouped in ranges for the study.  And the study shows that a patient’s age, tumor size and histology cannot be ignored in the exclusive favor of a DNA test.

The AD study looks at imaging for two molecules that are associated with dementia, tau and amyloid, which are detectable in brains with radionuclide PET scans, along with brain measurements made on MRI scans. The scans were scored as either positive or negative, not as continuous variables, although the clinical assessments were continuous (the graphs look like Mark Tobey art to me, but what do I know). The scintillations and measurements might be precise, but they don’t tell the whole story.

I grew up as a cancer clinician in the days when ordering a CT scan was a big deal, MRI was not available, and radionuclide scans were done down in “UNclear medicine.”  We were still figuring out the significance of leukemia karyotypes, before microarrays or any genomic testing.  But we were curing many kids with cancer with therapy that was developed by pioneers without sophisticated tools. In other words, diagnostic technology progresses and helps, but the most important impact for patients was therapeutic innovation.

Don’t get me wrong, I think that diagnostic work is really important, and it may also give clues to pathophysiology (i.e. is amyloid “necessary but not sufficient” to cause a patient’s problems?).  But many factors in cancer treatments emerged by analyzing clinical trials (in childhood ALL, simple things like age and total WBC were predictive of response, even before being able to order immunophenotype and molecular markers).

In some ways early cancer clinicians were “lumpers,”  grouping together patients because they couldn’t know any better, and are now better “splitters,” finding that more intricate subgroups can be identified which might change therapy decisions and improve outcomes.

So, how do we change the trajectory for AD patients? Well, the Alz Drug Discovery Foundation is one private group funding some exploratory new drug ideas.  The incidence numbers above suggest that there are subjects to be had, if a research clinician is a “lumper,”  and doesn’t have the wherewithall to pursue AT(N) testing, to allow a broader enrollment in an exploratory study. If something hits, maybe AT(N) testing could be done later (I may be wrong, but perhaps the study hints  that might be OK, in their mention of a limited number of repeated scans.) The most important factor might be leadership at the level of the National Institute of Aging, or the HHS/Nat’l Alz Project Act/Advisory Council, which seems to lack coherent concrete plans to develop therapy by their own deadline of 2025.

Time is always fleeting, so I guess late homework will have to be accepted, with a nice excuse note, but understand that people are waiting for answers.