BioMarkers & Diagnostic-Man vs AI

The World Health Organization (WHO),  has defined a biomarker as “any substance, structure, or process that can be measured in the body or its products an influence or predict the incidence of outcome or disease.” In Wikipedia.

Biomarkers are critical to the rational development of drugs and medical devices. Examples of biomarkers include everything from blood pressure and heart rate to basic metabolic studies and x-ray findings to complex histologic and genetic tests of blood and other tissues. Biomarkers are measurable and do not define how a person feels or functions. A diagnostic biomarker refers to a biological parameter that aids the diagnosis of a disease and may serve in determining disease progression and/or success of treatment. Biomarkers can be used to diagnose phenotyping, so to detect or confirm the presence of a disease, or to identify different diseases sub-types and even different habitats within a single lesion.

Those are:
-Blood Tests
-Brain Imaging
-Cerebrospinal Fluid
-Physiological Tests
-Saliva Tests
-Urine Tests
-Combination of Methods.

The complete blood count (CBC) is one of the most common blood tests. It is
often done as part of a routine checkup. This test measures many different
parts of your blood, including red blood cells, white blood cells, and
platelets. They help doctors check for certain diseases and conditions. They
also help check the function of your organs and show how well treatments are

But crucial point of having blood test is to interpretation of the numbers. Reference ranges (reference intervals) for blood tests are sets of values used by a health professional to interpret a set of medical test results from blood samples. Here starts problems and maybe misdiagnosis and misuse of medicines. References range may vary with age, sex, race, pregnancy, diet, use of prescribed or herbal drugs and stress, laboratories etc. Nothing In Life Is ‘One Size Fits All’ . For example, should on a lab  report only one reference range , say for HgA1c normal level is below 5.7%, a level of 5.7% to 6.4% indicates prediabetes, and a level of 6.5% or more indicates diabetes, and prescribe medicine for that, it might be a wrong decision. It may lead a misdiagnosis of diabetes in the elderly. This reference increasing to 6-6.5% for individuals aged 40–59 years while for people aged ≥60 years and might be 6.5-7 % even higher according to some researches. Reference intervals for men and women differ only slightly.

Another misleading point on some lab report  those limits either might be wrong or out-of-date. Very specific example of that here in Ankara on Public Health Lab report reference range for HDL written as between 40-60 and every time you get a lab report, you get your HDL level market as “our of range”. Whereas the higher HDL the better sign of health.  Other values also might not be consistent with the latest research outcomes.

Other than that physical fitness, exercise routine plays an important role and there are different numbers should be punched in for athletes’ categories. Mostly your age, sex and maybe weight and BMI be on the paper but not reference ranges for biomarkers. It is common to see many different counts due to hydration status, electrolyte balance (specifically Na+ & K+) may be outside the normal range. The lifestyle of an athlete at many levels is abnormal, and so too are their blood tests. For example It is well known that athletes have lower heart rates (bradycardia) than non-athletes. This is generally considered a healthy adaptation. But when you go for a check up or for any other reason and get your rate measured, you may be labelled as having a syndrome of something. Test followed by an exercise results very different numbers such as: Regular exercise causes an increase in the number of RBCs in the blood, a urine test may be positive for blood, anemia is a fairly common problem for runners, protein may be detected in the urine within 30 minutes of strenuous exercise, blood levels of CK- Creatine Kinase   are elevated following exercise, AST, one of the tests that evaluates liver function, levels will be elevated following exercise, blood sugar will be elevated. Besides these immediate effects long term differences requires the athletes be viewed in another windows of reference ranges.

Most of diagnostics are based on biomarkers. Even we assume that everything would be as is supposed to be and you have  %100 correct lab report, measurements, reference ranges, you are still way away from having a perfect diagnostic. Because these biomarkers, tens, hundreds should be interpreted correctly. Since there are lots of correlation of baseline biomarkers, it is very hard for a human being to see, evaluate and correlate these biomarkers and make a diagnosis.

Now In clinical practice efforts are already ongoing to apply AI to obtain new imaging data and improve the stepwise development of radiomics and validation of biomarkers, so instead simply human interpretation which is depend individual limited time and correlation capability AI will do the job much more faster and higher accuracy.

 We often see news like “120 pulmonologists from 16 European hospitals evaluated 50 cases with PFT and clinical information, resulting in 6000 independent interpretations. The AI software examined the same data. American Thoracic Society/European Respiratory Society guidelines were used as the gold standard for PFT pattern interpretation. The gold standard for diagnosis was derived from clinical history, PFT and all additional tests. The pattern recognition of PFTs by pulmonologists (senior 73%, junior 27%) matched the guidelines in 74.4±5.9% of the cases (range 56–88%).  Pulmonologists made correct diagnoses in 44.6±8.7% of the cases (range 24–62%) with a large interrater variability (κ=0.35). The AI-based software perfectly matched the PFT pattern interpretations (100%) and assigned a correct diagnosis in 82% of all cases (p<0.0001 for both measures).
Ref:Artificial intelligence outperforms pulmonologists in the interpretation of pulmonary function tests