In the current study, our stool and plasma study did not show a significant difference between a positive M2-PK test result and the distribution of age, sex, diabetes and family history of tumor in tumor- or polyp-bearing patients. There was only a significant difference between the results of M2-PK test in plasma samples of tumor-bearing subjects and normal subjects in terms of smoking (p = 0.011), although it was not seen in polyp-bearing subjects.
These findings were in consistent with the findings of U Haug et al, which reported that the subgroup of the ESTHER study did not differ from the whole ESTHER study population with respect to the distribution of age, sex, body mass index, smoking status and family history of CRC. However, current smokers showed more frequently increased levels of tumor M2-PK in stool compared to never and former smokers (p value=0.003)[17]. In a similar study, male and female groups showed no significant differences in age or fecal tumor M2-PK levels although a highly significant difference was found between the tumor M2-PK level for participants aged 20-49 years (median M2-PK of 0.66) and 50-79 years (median M2-PK of 0.086)[18]. Besides, in a study with 1082 participants (mean age 63 years, 50% females) the median tumor M2-PK level in the whole study population was 1.3 U/ml (0.3–3.3). Median tumor M2-PK levels did not alter by gender, but tended to be higher in older age groups (p value=0.002). Besides, the sensitivity and specificity did not vary by sex of stool samples. The specificity tended to be lower in older age groups (p value=0.001) but the sensitivity did not vary by age [19]. In another study, average serum M2-PK value among 158 normal individuals was 2.96 U/mL, which was not affected by gender or age [20]. The study of Mohamed El–Tantawy Ibrahim and colleagues revealed that there was no significant difference between patients with colon cancer and control groups considering the age and sex. Moreover, 32% of their patients were smokers compared to only 3.3% of the control group, which was statistically significant (p value less than 0.05) [3].
In our study, although in M2-PK plasma experiment the size of the tumor or polyp was statistically different in the tumor- or polyp-bearing patients in compare to controls, there was no difference between these groups in the M2-PK stool experiment. This was consistent with the study of Yogesh M. et al which reported that in patients undergoing colonoscopy 31 had adenomatous polyps, 21 had small adenomas (<10 mm) and 10 had large adenomas (>=10 mm). Median stool M2-PK in the small and large adenoma groups was 2.9 U/ml and 1.5 U/ml respectively, which was not statistically significant when compared with normal groups. M2-PK was reported positive in 25.8% of adenomas regardless of their sizes; however, FOBT seemed to be more associated with the size of the lesion [11]. In addition, in a similar study with 50 patients suffering from an adenomatous disease, 22 were found to have a single polyp greater than 1 cm in size. There was no significant difference in the M2-PK concentration detectable in the feces of patients with polyps less or above 1 or even the size of 5 cm [21].
In our study, ANOVA test revealed no significant difference in the location of tumor or polyp with a positive M2-PK test in either stool or plasma samples. However, Haug et al. showed that there was a statistically difference (p value <0.001) in tumor M2-PK levels in stool of ESTHER participants based on the location of the tumor. In their study with the cut-off value of 4 U⁄ml, overall sensitivity was 68% with a clear difference between colon cancer (85%) and rectum cancer (56%) [22].
In the present study, at the cut-off value of 15 U/ml, the test sensitivity for the plasma samples of tumor- and polyp-bearing groups were both 98%, specificity of 74%, positive predictive value of 75% and negative predictive value of 98%.
Besides, our results showed that the sensitivity of M2-PK test, at the cut-off value of 15 U/ml, was very high with a NPV of 98%, meaning that if the level of plasma M2-PK in an individual is determined less than 15 U/ml, the probability for a tumor or polyp is very low. The specificity of 74% for both groups (tumor and polyp-bearing groups) at the cut-off value of 15 U/ml indicated that this specificity is acceptable for laboratory testing, confirmed by the positive predictive value of 74%. At the cut-off value of 4 U/ml, the test sensitivity, specificity, PPV and NPV for the stool samples of tumor-bearing group was 100%, 68%, 52.7% and 100%, respectively.
Chi-square test was used to compare the normal and abnormal levels of M2-PK in the stool samples of the normal and polyp-bearing groups, which showed a significant difference between the two groups (p= 0.0001). At the cut-off value of 4 U/ml, the test sensitivity for the stool samples of polyp-bearing groups was 87%, specificity was 68%, PPV was 65% and NPV was 88%.
The sensitivity of fecal M2-PK test was higher in tumor-bearing group (100%) than in polyp-bearing group (87%). Besides, NPV was 100% in tumor-bearing group, meaning that if the level of fecal M2-PK of an individual is determined less than 4 U/ml, the probability for a tumor is almost zero. In contrast, regarding the low PPV of M2-PK test for detecting tumor and polyp in stool, any result higher than 4 U/ml can be false positive indicating a low specificity of the test. Other researchers have shown that the sensitivity of M2-PK measurement in the feces was 77.9% for CRC and specificity ranged from 74.3% to 83.3. Overall sensitivity for adenomas was 45.9%, raising to 61.1% for adenomas larger than 1 cm [16]. In another study, the pooled sensitivity and specificity of fecal M2-PK for the diagnosis of CRC calculated by the bivariate model were 79% and 81%, respectively and the summary PPV and NPV were 74% and 86%, respectively [23]. In a study performed by Kumar et al, at a diagnostic cut-off value of 15 U/ml for plasma tumor M2-pyruvat kinase, sensitivity, specificity, PPV and NPV were 57.3, 89, 85.7 and 64.8%, respectively. In CRC, fecal tumor M2-PK had a sensitivity of 73-92% at a cut-off value of 4 U/ml in compared to 50% sensitivity for Guaiac fecal test [24]. In a multi-center study on 317 subjects with a cut-off value of 4 U/ml, fecal M2-PK assay had a sensitivity, specificity, PPV and NPV of 81.1, 71.1, 61.9, and 86.7% respectively to detect CRC [25]
Besides, with 328 patients according to ROC, the tumor M2-PK cut-off level was 4.00 U/mL (the sensitivity, specificity, PPV, and NPV were 71.4%, 71.0%, 73.5%, and 94.4%, respectively) [20]. In a study by Hisham K. Dabbous et al, M2-PK was the most sensitive and specific test in differentiating CRC from control subjects in fecal samples with sensitivity and specificity of 75%, and 100%, respectively [14].
In the present study, ROC curve was used to determine the best cut-off value for tumor M2-PK test. Based on our results, in order to have the best sensitivity and specificity for tumor detection in plasma specimens, a cut-off value >25 U/ml was recommended, in which the test sensitivity and specificity was 90.9% and 91.3%, respectively. Besides, for polyp detection, a cut-off value >19 U/ml was suggested, in which the sensitivity of the test was 96.3% and the specificity was 85.5%. AUC of polyp data was 0.95 and for tumor data was 0.975 in plasma, which confirms that the overall discriminatory power of the test is high.
Furthermore, in order to have the best sensitivity and specificity for tumor detection in stool samples, a cutoff value >8 U/ml was recommended, in which the test sensitivity is 100% and the specificity is 85.6%. For polyp detection, a cut-off value >4.8 U/ml was suggested, in which the sensitivity of the test was 81.6% and the specificity was 74.8%. AUC of polyp data was 0.834 and for tumor data was 0.969 in stool that confirmed the overall discriminatory.
Therefore, a cut-off range of 4.8-8 U/ml in stool samples can detect polyp and a cut-off value > 8 U/ml can detect tumor. In addition, a cut-off range of 19-25 in plasma samples can detect polyp and a cut-off value > 25 can detect tumor.
In our study, the specimen size of tumor was not the original size because it was based on the biopsy sample. The same situation applied for polyp samples except for the time when polypectomy was performed, having an original. The specimen size of tumor might affect the analysis related to the M2-PK level and sample size of tumor or polyp in our study.