In this translational study, we evaluated the diagnostic utility of cfDNA fragments, DNA integrity index and mesothelin in the 75 matched pleural effusion and serum samples and found that pleural fluid DNA integrity index is higher in MPE (mesothelioma and other malignancies) compared to benign effusions. Importantly DNA integrity index had diagnostic utility in detecting MPE even in cytology negative effusions. In contrast, the diagnostic utility of pleural effusion and serum mesothelin was limited to mesothelioma but not non-mesothelioma MPEs.
It has been widely reported that total cfDNA in the plasma and serum is higher in cancer patients compared to healthy controls and may serve as a potential diagnostic biomarker for solid organ malignancies [9
]. However the diagnostic utility of total cfDNA in plasma and serum for lung cancer has been questioned since the total cfDNA in of patients with non-malignant lung disease is not significantly different from that of lung cancer patients [18
]. Additionally there is considerable variability in the reported values of total cfDNA in cancer patients, patients with non-malignant illnesses and healthy controls, limiting utility as a diagnostic biomarker [9
]. Previous reports have demonstrated that cfDNA is 4-6 fold higher in serum compared to plasma [23
] and maybe a better biological specimen to screen for cfDNA in malignancy [13
]. Hence we used matched serum as a reference comparator to pleural fluid in evaluating the utility of cfDNA and DNA integrity index in diagnosing malignancy. An alternative approach to distinguish malignant from benign effusions has been to evaluate the utility of quantitative and qualitative tumour-specific alterations, such as microsatellite alterations in effusion DNA [24
]. Economidou et al. studied patients with malignant (n=26) and benign (n=22) effusions and found that microsatellite instability and loss of heterozygosity in DNA from pleural fluid and blood were not diagnostically useful [27
]. Conversely, Benlloch et al [24
] and Chan et al [25
] demonstrated that pleural effusion total DNA was significantly higher in MPE compared to benign effusions. However, in the study by Benlloch et al, all parapneumonic effusions were excluded from analysis and MPEs were compared predominantly to transudate effusions [24
]. Chan et al. did not exclude parapneumonic effusions and found that MPEs had higher total DNA compared to both benign transudate effusions and infective effusions [25
]. However they also found that while MPE total DNA in pleural fluid is significantly higher compared to transudate effusions (p<0.0001), it was less robustly significant compared to infective effusions (p=0.048) [25
]. Additionally, in the group of infective effusions, the majority (82%) was due to tuberculosis, a subgroup that is not represented in our study.
Recently several studies have showed that in comparison to healthy controls, serum DNA integrity index is higher in breast tumours [29
] and prostate tumours [30
]. and plasma of head and neck tumours [31
], naso-pharyngeal carcinoma [32
] and rectal cancer [33
]. In contrast we did not find a difference in serum DNA integrity index between subjects with MPE and benign effusions. This is most likely because of our subjects with inflammatory pleuritis and parapneumonic effusions accounted for 35% and 17% of the benign effusion cohort. Our findings are consistent with the findings of Schmidt et al, who also did not find a difference in the serum DNA integrity index of lung cancer patients compared to those with non-malignant lung disease [34
]. This is most likely due to the release of increased quantities of DNA into the blood from both apoptotic and necrotic cells in severe inflammatory processes [35
To our knowledge, only one other study by Salani et al, has evaluated the diagnostic utility of DNA integrity index in effusions [26
]. They measured the integrity index of cyclin E, a gene frequently amplified in ovarian carcinoma in ascites and pleural effusions. This study demonstrated that cyclin E assay was specific to ovarian carcinomas, thereby limiting its usefulness as a diagnostic test for all MPEs. In our study we used ALU repeats to measure pleural effusion DNA integrity index, since ALU repeats are ubiquitous and the ALU qPCR assay has previously been validated in several different malignancies [13
Cytology remains the diagnostic standard for evaluating pleural effusion samples. However it is can be difficult to establish a diagnosis of MPE in subjects with cytology-negative effusions, since in such circumstances diagnosis requires invasive pleural biopsy tests, such as thoracoscopy. We found that when combined with cytology, pleural fluid DNA integrity index substantially increased the ability to distinguish benign effusions from MPE (sensitivity 81% vs. 55%) and mesothelioma (sensitivity. 81% vs. 31%). Importantly, almost a third of false positive pleural fluid DNA integrity index results were due to parapneumonic effusions. Additionally, we showed that in cytology negative effusions, elevated pleural fluid DNA integrity index had 81% PPV. Hence pleural fluid DNA integrity index provides valuable additional information to pleural fluid cytology, particularly in subjects with either inconclusive or “suspicious” cytology results. This may have clinical implications since an elevated pleural fluid DNA integrity index in a cytology-negative subjects should be prioritized to undergo thoracoscopy.
Our study included subjects with mesothelioma (31% of subjects with MPE), a subgroup that was present in only 0-3% of other studies that have studied total DNA in effusions [24
]. It is now important to identify accurate diagnostic biomarkers for mesothelioma since it is increasing in incidence, with a peak predicted to occur between 2014 to 2021 [37
]. Consistent with previous studies, we found that pleural effusion and serum mesothelin has diagnostic utility for mesothelioma [7
]. However a weakness of our findings is that serum mesothelin results could not be obtained in 13% (10/75) of study subjects. Nonetheless, our data demonstrate that pleural fluid DNA integrity index is comparable in sensitivity to pleural fluid and serum mesothelin in diagnosing mesothelioma.
There are certain limitations to consider before the results of our study can be applied to other patient populations. Firstly, while our study has found that DNA integrity has utility in distinguishing MPEs from benign effusions, it should be noted the method used for measuring DNA integrity in our study is one of several others that have been reported in the literature [12
]. In this study, we have used the ratio of the longer DNA fragments to the shorter DNA fragments to measure DNA integrity (ALU 247/115), as described by Umetani et al [13
]. This methodology has since been used by other investigators to measure DNA integrity in the serum of breast cancer patients [36
] and plasma of rectal cancer patients [33
]. Most recently, Agostini et al found that DNA integrity (ALU 247/115) in the plasma of patients with rectal adenocarcinoma (n=67, median, IQR25-75 1.1 (0.7-1.9)) compared to healthy control subjects (n=35, median, IQR25-75; 0.1 (0.0 – 0.4)) [33
]. The findings of DNA integrity in malignancy are similar to our study where the median DNA integrity in serum 0.9 (0.5-1.3). Interestingly, the serum DNA integrity index in our cohort of subjects with benign effusions was also relatively high and not significantly different from subjects with MPEs (median, IQR25-75; 0.9, 0.5-1.0). This is not unexpected since our ‘control’ population were not normal healthy volunteers but subjects with non-malignant pulmonary diseases. Secondly the different studies have not only used other methods to measure DNA integrity, they have also used other protocols to process and extract DNA from blood [12
]. Recently Fleischhacker et al found that significantly different amounts of absolute DNA values were obtained from plasma using different DNA isolation methods [39
]. Indeed Jung et al, note that the considerable heterogeneity in preanalytical and analytical factors considerably determine the interpretation of cfDNA and DNA integrity studies in malignancy, thereby limiting the translation of this test into clinical practice [12
]. Hence, despite extensive research into the utility of cfDNA and DNA integrity index as diagnostic and prognostic biomarkers in malignancy, the methodological discrepancies have dictated that the tests have remained as research tools. Thirdly, another limitation of our study relates to the relatively small number of subjects evaluated and the casemix of pleural effusion diagnoses, particularly the relatively high proportion of mesothelioma subjects (31%), derived at a single Australian thoracic tertiary referral centre. It is well recognised that the distribution of aetiologies of pleural effusions (including MPEs) vary considerably according to geographic location [40
]. Further validation will be needed to address issues of generalisability and applicability. Finally, another restriction of our study is that the DNA extracted from a mixture of various types of cells present in pleural fluid (tumour cells, mesothelial cells, white blood cells). The influence of non-tumour cells on DNA fragments is currently not known and will also need to be evaluated. For these reasons, our findings should be interpreted as a pilot study, albeit with results suggesting that pleural fluid DNA integrity index is a promising diagnostic biomarker for MPEs. The results require confirmation in prospectively conducted multicentre studies with sufficient attention paid to the methodological variations common in cfDNA and DNA integrity index research before they can be incorporated into clinical decision-making algorithms.