Comparison Of X-ray Diagnosis Of Silicosis With Autopsy Data

Comparison Of X-ray Diagnosis Of Silicosis With Autopsy Data

Z Ndlovu

NIOH and University of Calgary, Canada

Project Summary:

The Occupational Disease in Mines and Works Act of 1973 (ODMWA) makes provision for the lifelong medical surveillance and evaluation of both active and former miners for possible compensable occupational lung disease. Employed miners undergo regular surveillance and ex-miners can be examined in designated provincial public hospitals. In addition, statutory autopsies are carried out at the NIOH on these miners to determine the presence of occupational lung diseases for compensation purposes as outlined in the Act.

The burden of compensable occupational lung disease in South African miners is high. South African gold miners are exposed to dust with a high content of crystalline silica and silicosis is prevalent among these miners. Churchyard and coworkers (2004) showed the prevalence of silicosis to be 18-20% in active black gold miners. The prevalence of silicosis was 22-37% in black, gold ex-miners (Trapido, 1998) and 54% in white ex-miners (Murray, 2005). The autopsy data for 2005 show an overall silicosis rate of 223 per thousand and 295 per thousand in gold miners (Ndlovu et al., 2006).

The assessment of the presence of silicosis in dust exposed workers is based mainly on the presence of rounded opacities on radiographs. Two studies in South African gold miners have investigated the correlation between the radiological and pathological diagnosis of silicosis. The reliability of the radiological readings for the diagnosis of the presence of silicosis was evaluated using sensitivity, specificity and predictive value statistics. Sensitivity is the percentage of positive radiological readings in cases where silicotic nodules are found in the lungs. Specificity is defined as the percentage of negative radiological readings in cases with no or insignificant numbers of silicotic nodules found at autopsy. It was assumed that the sensitivity and specificity measurements were dependent on the readers’ interpretation of the X-rays.

Hnizdo and co-workers (1993) found sensitivities of 0.39, 0.37 and 0.24 and specificities of 0.99, 0.97 and 0.98 for the three X-ray readers who participated in the study. Corbett and co-workers (1999) calculated sensitivity of 71% and specificity of 96%. Both these studies showed a high proportion of false negatives i.e. silicosis at autopsy but not on radiology and more false negatives than false positives (approx 66% and 29%, respectively). Sensitivity improved with increasing degree of silicosis at autopsy.

The specificities in both the studies described above were high. Using these data, one can expect between 1 and 4% of false positives. The anomalies could have been due to misinterpretation of radiographs due to the presence of other conditions in the lung which have presentations similar to those of silicosis. There are consequences attached to the inaccurate diagnoses of silicosis. The cost associated with false positive radiological diagnosis is the compensation paid to miners who are not diseased. Although the numbers of false positives is low, this study may yield information that could be useful to radiologists. The study may show that there are certain diseases that mimic silicosis radiologically e.g. sarcoid.

Contact Person:
Ms. Z. Ndlovu (zodwa.ndlovu@nioh.nhls.ac.za)