MIMIT seminar series - prostate cancer

Histopathology Made Simple: Developing Spectroscopic Methods of Grading Prostate Tissue –
Dr Peter Gardner

Cancer occurs when the balance between oncogenes and tumour suppressor genes is altered, leading to cell metastasis. With no known cure, it is important to develop existing methodologies in order to aid early diagnosis and offer effective treatment so survival rates are increased. At 23%, prostate cancer it is one of the most common types of cancer and affects 32,000 men annually in the UK. It is slow growing and it may take years for the symptoms to appear.

Diagnosis methods include a rectal exam and blood tests to check prostate specific antigen (PSA) levels. Information can also be obtained from needle biopsies, X-rays, ultrasounds and magnetic resonance imaging (MRI) scans. Usually needle biopsy samples are microscopically analysed and graded using the Gleason grading system by the pathologist. This method is highly subjective and previous studies have shown that results vary significantly between pathologists. The aim is to produce an objective method of diagnosis which has an accurate and reliable outcome that is no longer dependent on the pathologist.

The group proposed the use of Fourier transform infrared spectroscopy (FTIR) to detect malignancy in tissues by assessing tissue architecture and biochemical activity. The FTIR data was combined with a linear discriminate analysis (LDA) to create an operator independent diagnostic algorithm. Previous papers have shown that the 1030cm-1 band to be indicative of glycogen and the group’s work showed that the ratio between absorption bands 1030cm-1 and 1080cm-1 may be used as metabolic markers for cancer. Using the FTIR-LDA model, to test paraffin embedded samples with a known Gleason score, produced a good correlation using the 3 band gleason grading system.

It was also found that the biochemical activity changed in cells before the tissue structure suggesting that the FTIR-LDA model may be able to use biochemical activity as a marker as opposed to tissue structure to grade cancer. The FTIR-LDA model is more suitable as it is an objective method that doesn’t require the use of dyes or stains, is operator independent and has a good resolution.

The work done so far is promising and the group are looking at a number of future possibilities. Firstly the automation of the process would help to simplify the process and also cut down the time scale required for diagnosis in comparison to the fixation process currently used. The use of the biochemical activity as a marker is also another possibility which would allow earlier detection possibility with better sensitivity and specificity. There is also the idea of combining the grading method with an imaging process, similar to face recognition software, in order to identify tissue structure to determine the disease stage. A final point would be to improve the speed of the equipment in order to improve the speed of data acquisition.

In conclusion, the model is a good method but requires much more work before it can be put into practice. Legal and ethical issues need to be considered to produce a large set of biopsy samples, from which an appropriate and reliable reference standard can be created for comparison. The procedure must also be cheap and easy to use and the technology perfected so that people will make the switch from the traditional subjective methods to more modern methods for cancer diagnosis.