Evaluation of GenoType MTBDRplus by Use of Extracted DNA from Formalin-Fixed Paraffin-Embedded Specimens

LETTER

The number of cases of extrapulmonary Mycobacterium tuberculosis complex infection increased during the past waves of arrival of refugees in Europe (1–3). As extrapulmonary disease could be misidentified as a space-occupying malignancy, physicians might request only a histopathological examination and no microbiological culture. For that purpose, specimens are routinely formalin fixed and paraffin embedded (FFPE). We frequently experienced a diagnostic dilemma if acid-fast bacilli were seen, because performance of an ex post microbiological culture for susceptibility testing from FFPE specimens is not feasible. However, antimicrobial susceptibility testing is warranted, particularly in populations with a high prevalence of multidrug-resistant tuberculosis (4). Several nucleic acid amplification tests are available for the identification of M. tuberculosis and rifampin (± isoniazid) resistance, but these tests are licensed only for culture and pulmonary/sputum specimens (5). Here, we evaluate a line probe assay as a rescue approach for the identification of rifampin and isoniazid resistance in M. tuberculosis using DNA extracted from FFPE specimens.

Histopathological samples (n = 14) (Fig. 1) were prospectively collected from routine diagnostic procedures at the University Hospital Münster, Münster, Germany, between 2015 and 2017. They were included in the study if either (i) acid-fast bacilli (Ziehl-Neelsen stain) or (ii) M. tuberculosis complex-specific gene segments were detected using the MYCO^Direct 1.7 assay (Chipron, Berlin, Germany). A detailed description of the methods is given in the supplemental material.

FIG 1.

Study flow. In total, 14 tissue samples were positive for Mycobacterium tuberculosis (MYCO^Direct array) using DNA extracted from formalin-fixed paraffin-embedded specimens. These DNA samples were used for the GenoType MTBDRplus test (Hain) to identify resistance to rifampin and isoniazid. In parallel, mycobacterial cultures and susceptibility testing were done for 12 samples to assess the agreement between GenoType MTBDRplus susceptibility results and those obtained with the gold standard (culture).

All samples were positive in the MYCO^Direct assay, but only 30.8% (4/13; not done, n = 1) showed acid-fast bacilli (Table 1). Others have reported similar low (9.4% to 25%) detection rates of microscopy compared to nucleic acid amplification assays as well (7, 8).

TABLE 1.

Comparison of GenoType MTBDRplus using DNA extracted from formalin-fixed paraffin-embedded specimens and microbiological culture^a

Specimen no.	Specimen source	AFB in microscopy	DNA quality (largest amplifiable fragment [bp])b	DNA result from formalin-fixed paraffin-embedded specimen (GenoType MTBDRplus)				Microbiological culture result
				Concn [ng/μl]	Identification	Rifampin	Isoniazid	Identification	Rifampin	Isoniazid
1	Vertebra	Negative	600	16.2	MTB complex	S	S	M. tuberculosis	S	S
2	Lymph node	Positive	400	6.8	Nonevaluable	NA	NA	M. tuberculosis	S	S
3	Liver	Negative	400	8.7	MTB complex	S	S	M. tuberculosis	S	S
4	Os sacrum	Negative	300	6.37	Nonevaluable	NA	NA	No growth	NA	NA
5	Lymph node	Negative	100	131	MTB complex	S	S	M. tuberculosis	S	S
6	Os sacrum	Negative	200	5.8	Negative	NA	NA	No specimen	NA	NA
7	Mamma	Positive	400	54.9	MTB complex	S	S	No growth	NA	NA
8	Lung	Positive	300	148	MTB complex	S	S	M. tuberculosis	S	S
9	Lymph node	Negative	400	431	MTB complex	S	S	M. tuberculosis	S	S
10	Lymph node	Negative	400	351	MTB complex	S	S	M. tuberculosis	S	S
11	Lung	Not done	400	118	MTB complex	S	R	No growth	NA	NA
12	Lymph node	Negative	400	107	Negative	NA	NA	M. bovis	S	S
13	Lung	Negative	300	68.6	MTB complex	S	S	No growth	NA	NA
14	Kidney	Positive	100	2.94	Nonevaluable	NA	NA	Not done	Not done	Not done

^aAll samples were positive in the MYCO^Direct assay (PCR and hybridization). AFB, acid-fast bacilli (Ziehl-Neelsen stain); NA, not applicable; R, resistant; S, sensitive.

^bData are from reference 6.

We extracted DNA from all M. tuberculosis complex-positive FFPE specimens (median concentration, 61.75 ng/μl; range, 2.9 to 431 ng/μl). The PCR product sizes of amplifiable DNA fragments in a multiplex DNA quality control PCR ranged between 100 and 600 bp (Table 1) (6).

All samples were tested with the GenoType MTBDRplus line probe assay. In total, 11 (79%) tests were evaluable; three (21%) tests were not valid because either the hybridization zone of M. tuberculosis-specific amplicons or the rpoB locus or the inhA locus or both were not developed (see the instructions for use provided by the manufacturer). The median DNA concentration of evaluable tests (107 ng/μl; range, 5.8 to 431) was significantly higher than in nonevaluable tests (6.4 ng/μl; range, 2.9 to 6.8; P = 0.02 [Wilcoxon rank sum test]) (Fig. 2). The concentration of DNA, consisting of both human and mycobacterial DNA, seems to be a suitable surrogate to predict evaluable test results.

FIG 2.

Test performance according to the DNA concentration. DNA from formalin-fixed paraffin-embedded specimens was used to perform the GenoType MTBDRplus test to detect Mycobacterium tuberculosis and resistance to rifampin and isoniazid. The scatterplot compares the DNA concentrations of evaluable and nonevaluable test results. The bars show the upper and lower quartiles as well as the median (thick bar).

In 9 of 11 evaluable tests, we detected M. tuberculosis; all samples were susceptible to rifampin according to GenoType MTBDRplus testing. Isoniazid resistance was detected in one sample (Table 1).

In parallel, non-formalin-fixed microbiological specimens were obtained from 12 of the 14 patients and were sent for mycobacterial culture (Fig. 1 and Table 1) (9). The median time to positive culture of M. tuberculosis (n = 7) and Mycobacterium bovis (n = 1) was 15.5 days (range, 8 to 58 days). The line probe assay and microbiological culture yielded discordant results in five cases as follows: the GenoType MTBDRplus test result was negative/nonevaluable despite microbiological growth (n = 2), or the GenoType MTBDRplus test detected M. tuberculosis complex despite no growth (n = 3; Table 1). In general, the susceptibility of rifampin and isoniazid as detected by GenoType MTBDRplus was always confirmed by the culture-based antimicrobial susceptibility testing.

A similar study on extrapulmonary tuberculosis showed that the sensitivity and specificity of the GenoType MTBDRplus assay using DNA from FFPE tissue samples were 60% and 85.7%, respectively, compared to a nonmicrobiological case definition (8).

In conclusion, DNA from FFPE specimens can be used in GenoType MTBDRplus assays to detect isoniazid and rifampin resistance. This provides guidance to select the best antimycobacterium therapy if culture-based susceptibility testing is not feasible.