Pleural Mesothelioma

Computer-aided diagnosis of early-stage pleural mesothelioma: automatic detection and quantitative assessment of pleural thickenings in thoracic CT data

The background

For almost a century, asbestos was imported and used in the industrialised countries of Western Europe and North America before the fibrotic and carcinogenic danger of inhaling asbestos fibres was recognised. In fact, malignant pleural mesothelioma, normally a rare tumour, was found a thousand times more frequently in people regularly exposed to asbestos in their daily work than in the general population. Today it is statistically documented that 70-90% of malignant pleural mesotheliomas are attributable to asbestos exposure. Almost three decades passed before legal bans were introduced, in Germany for example in 1993. There are different kinds of asbestos, distinguished by the shape of their fibres. The narrow, needle-like kind (amphiboles such as crocidolite and amosite, so-called blue asbestos) was classified as more carcinogenic than the curled, flexible kind (chrysotile, so-called white asbestos), which is more biodegradable and therefore less dangerous. Both kinds were widely used across industries. Because of post-war industrialisation, asbestos consumption in Germany, for example, reached its peak during the late 1970s. Owing to its long latency period of 35 years on average (between 10 and 60 years) after asbestos exposure, the peak in malignant pleural mesothelioma cases and subsequent deaths in Germany is reached during the 2010s. Worldwide, an increase to at least one million deaths is expected over the next 30 years. A screening programme for groups of people exposed to asbestos is currently in place in Germany. For the high-risk group this screening includes a thoracic CT scan as a non-invasive diagnostic procedure. Diagnosing physicians usually have to examine more than 80 thoracic CT slices visually in order to find all kinds of pleural or pericardial thickening. After a diagnosis of malignant pleural mesothelioma, survival of between 4 and 18 months can be expected without treatment. If pleural mesothelioma is diagnosed at an early stage, however, a pleurectomy together with perioperative treatment can reduce morbidity and thus delay mortality. An efficient and reliable diagnosis of early-stage pleural mesothelioma is therefore a key factor in containing the consequences of the expected peak in malignant pleural mesothelioma cases.

Thorakoskopie von pleuralen Mesothelioma	The expected increase in pleural mesothelioma morbidity in Germany, with a peak of at most 1000 cases in 2017.

The vision

The typical non-invasive clinical diagnosis of pleural mesothelioma is usually based on axial thoracic CT scans. Depending on the slice thickness, the number of images varies between 80 slices of 5 mm thickness and about 700 slices of 0.5 mm thickness. Diagnosing physicians examine each slice visually on a workstation in order to find pleural thickenings. Lung nodules are an incidental finding. The findings are documented in a standardised form containing data such as size, position and growth rate. This assessment method is very time-consuming and takes about 20 to 30 minutes per data set. It has also frequently been considered subjective. Indeed, studies have shown that the associated complexity of tumour measurement hinders accurate assessment. Differences in diagnostic results between different readers have also been found (inter-reader variability). Increasing the accuracy of the localisation and the topological information of these rather small image regions by subjective visual assessment would require a longer reading time for each data set. Because of the growing number of examinations and the high workload of the physicians concerned, however, that solution is not viable. Consequently, a method is needed that provides a more accurate assessment of early-stage pleural mesothelioma and that is also reliable, consistent and reproducible.

Unser Ziel

The aim of this work is to develop a computer system that automatically identifies pleural thickenings in axial thoracic CT images and quantifies them. When malignant pleural mesothelioma is suspected, it is very important to determine whether a new neoplasm is present, since the earlier malignant pleural mesothelioma is found, the longer the survival that asbestos patients can expect. When treatment follows a positive diagnosis of pleural mesothelioma, accurate and reliable observation of the development over the relevant period, such as the growth rate, is essential.

Our computer system was tested on the real data sets made available to us. The test results proved reliable and reproducible.

Graphical user interface of the PleuraDatPlus system
Determination of candidates for pleural thickening.	Measurement of the topological properties.

Our system is currently being integrated into clinical routine use.

Kontakt:	Peter Faltin, Tel.: +49 (241) 80-27865
Kooperationspartner:	Univ.-Prof. Dr. med. Thomas Kraus
	Institute for Occupational and Social Medicine,
	University Hospital Aachen, AöR,
	– Faculty of Medicine, RWTH Aachen University –
Funding:	Deutsche Gesetzliche Unfallversicherung (DGUV)
	Projekt-Nr. FF-FB0148
Letzte Aktualisierung:	02.03.2011

 

---