| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
1 Department of Epidemiology German Centre for Research on Ageing Heidelberg, Germany
2 Division of Clinical Epidemiology and Aging Research German Cancer Research Center Heidelberg, Germany
3 Department of General Surgery University Hospital of Heidelberg Heidelberg, Germany
aAddress correspondence to this author at: Division of Clinical Epidemiology and Aging Research, German Cancer Research Center, Bergheimer Strasse 20, D-69115 Heidelberg, Germany. Fax 49-6221-5481-42; e-mail uhaug{at}dkfz-heidelberg.de.
To the Editor:
Colorectal cancer (CRC) remains the third most common malignancy worldwide (1)(2). To improve noninvasive screening for CRC, various stool tests have been described based on tumor-associated markers (3). Among these is a test for fecal tumor M2 pyruvate kinase (M2-PK) activity (4). Tumor M2-PK, an isoform of PK, is found in proliferating tissues such as tumor cells (5). To date, the performance of this test has been evaluated only in small-scale investigations (4)(6). Regarding large-scale applications, the stability of tumor M2-PK, which would affect the handling of stool samples, is a critical issue. We investigated the average stability of tumor M2-PK in stool at room temperature to estimate the potential impact of its degradation on the sensitivity and specificity of the test.
We collected stool specimens before bowel preparation for surgery from 20 patients with histologically confirmed CRC. Samples were kept at room temperature for 5 days. Immediately after sample collection and on each of the following days, the amount needed for duplicate determinations of the tumor M2-PK activity was taken and stored at –20 °C until analysis. Samples were analyzed blinded to patient identity and day of sampling.
Tumor M2-PK was measured by ELISA (ScheBo® Biotech AG). The lower detection limit was 2 kU/L.
For each measurement point, the mean tumor M2-PK activity was calculated as the mean of duplicate determinations. For further analyses, only patients with an initial tumor M2-PK activity >4 kU/L, the proposed cutoff, were included. The values for each patient were expressed as a proportion of the initial tumor M2-PK activity. Linear regression, with log(tumor M2-PK activity) as dependent variable, was performed to obtain the relative activity for each day of storage at room temperature over all included patients.
To estimate the potential impact of degradation of tumor M2-PK on the diagnostic accuracy of the test, results of stability testing were used in combination with fecal tumor M2-PK activities measured in 65 CRC patients (26 with colon cancer and 39 with rectal cancer) and in 917 unselected older adults. Stool samples from CRC patients were frozen immediately after specimen collection, whereas stool samples from unselected older adults were collected by mail in the context of the ESTHER study (7) and frozen at –80 °C on receipt. The study was approved by local and state Ethics Committees.
The theoretical tumor M2-PK activity after n days of storage at room temperature was estimated as the relative activity after n days multiplied by the initial activity in CRC patients (to determine sensitivity) and of ESTHER study participants (to determine specificity). Given that stool samples from ESTHER study participants were mailed (i.e., stored at room temperature for 1–4 days), the equation was first transformed to calculate initial activities before proceeding as described. Finally, the sensitivity and specificity of the test were calculated for each day of storage at room temperature, using the cutoff of 4 kU/L.
Overall, 13 patients fulfilled the inclusion criteria (initial tumor M2-PK activity 
4 kU/L). Their initial mean tumor M2-PK activity was 13.1 kU/L. Regression analysis as described yielded a degradation rate of 18% per day (95% confidence interval, 13%–23%). After 1, 2, 3, 4, and 5 days of storage at room temperature, the relative activities were 73%, 60%, 49%, 40%, and 32%, respectively.
The Spearman correlation coefficient, used as an indicator of test-retest repeatability (8), was 0.93 (P <0.001).
The potential impact of tumor M2-PK degradation on test performance characteristics is shown in Table 1
. In CRC patients, the median (interquartile range) initial activity was 8.6 (2.8–18.0) kU/L, yielding a sensitivity of 68%. Initial specificity was 73% [based on initial activities calculated for ESTHER study participants: median (interquartile range), <2 (<2 to 4.7) kU/L]. After 1 day of storage at room temperature, sensitivity decreased to 62%, whereas specificity increased to 77%. During the following days, the loss of sensitivity also exceeded the gain in specificity.
|
We conclude that the handling of stool samples affects performance characteristics of the tumor M2-PK stool test. The size of the effect may vary in individuals, as our data allowed only indirect estimation of the impact of degradation by combining the results from different study samples. Another potential limitation concerns estimates of specificity. The group of unselected older adults serving as the control group did not undergo colonoscopy, and undiagnosed CRC among controls cannot definitely be excluded. However, their proportion would be expected to be very small (<1%) and unlikely to have led to appreciable bias (9)(10). Although assessed only for the tumor M2-PK stool test, the stability issues addressed in our study may also have important practical relevance for other stool tests intended to detect CRC in the screening setting, regarding the question of whether mailing of samples is possible without loss of information.
Footnotes
1 U.H. and M.N.W. contributed equally to this work. 
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
