On 2016 May 15, Rima Obeid commented:

On comparing the performance of single B12 markers (comparators) to detect cB12 indicator (separator) (Figure 1. ROC curves showing the diagnostic value of vitamin B-12, MMA and HC in predicting cB12: The study contained important results on possible disagreement between B-12 markers in patients with cancer. However, the diagnostic accuracy of B12, methylmalonic acid and Homocysteine in predicting cB12 (separator or outcome) by using ROC analyses and the interpretation of this approach need further evaluation. The separator (cB12) is calculated from the three B12 markers. Thus, cB12 cannot be used to define performance of the single components used for calculating it. Moreover, the three markers are not independent, but affected by each other. Therefore, the 3 individual markers that are indeed components of the new variable (cB12) participate to different degrees in determining the cB12 and they cannot be used in ROC analyses to predict sensitivity and specificity or best performance. ROC analyses aim at testing a 'new test or a new method' against established ones. A diagnosis that is based on using this 'new test under evaluation' should be made using criteria independent from the established tests being tested against this new test.

On 2016 May 15, Rima Obeid commented:

On comparing the performance of single B12 markers (comparators) to detect cB12 indicator (separator) (Figure 1. ROC curves showing the diagnostic value of vitamin B-12, MMA and HC in predicting cB12: The study contained important results on possible disagreement between B-12 markers in patients with cancer. However, the diagnostic accuracy of B12, methylmalonic acid and Homocysteine in predicting cB12 (separator or outcome) by using ROC analyses and the interpretation of this approach need further evaluation. The separator (cB12) is calculated from the three B12 markers. Thus, cB12 cannot be used to define performance of the single components used for calculating it. Moreover, the three markers are not independent, but affected by each other. Therefore, the 3 individual markers that are indeed components of the new variable (cB12) participate to different degrees in determining the cB12 and they cannot be used in ROC analyses to predict sensitivity and specificity or best performance. ROC analyses aim at testing a 'new test or a new method' against established ones. A diagnosis that is based on using this 'new test under evaluation' should be made using criteria independent from the established tests being tested against this new test.