AN INTELLIGENT SCREENING SYSTEM FOR CERVICAL CANCER

Cervical cancer is a worldwide problem affecting women. Malaysia has a population of 11.34 million women aged above 14 who are at risk of developing cervical cancer. According to the statistics last updated on 14 July 2014, every year, 2145 women are diagnosed with cervical cancer and 621 die from the disease. Cervical cancer is the second most frequent cancer among women. In actual fact, cervical cancer is not a death sentence as it is preventable and treatable if detected at an early stage. The mortality rate due to cervical cancer has been reduced tremendously as a consequence of widespread Pap-smear screening. Nevertheless, Pap-smear screening tests do not always yield accurate results as it is prone to human error such as misdiagnosis due to inexperience and the issue of human fatigue.

The International Agency for Research on Cancer (IARC), the specialized cancer agency of the World Health Organization, released the latest data on worldwide cancer incidence, mortality, and prevalence on 12 December 2013. According to the updated version of IARC's online database, GLOBOCAN 2012, cervical cancer is the third most common cancer among women worldwide, with an estimated 83,195 new cases and 35,673 deaths in 2012. The estimation of cervical cancer deaths annually in Malaysia was about 621 in 2012. Cervical cancer is the fifth leading cause of female cancer-related deaths in Malaysia and is the fourth leading cause of cancer-related deaths in women aged 15 to 44 in Malaysia, according to data updated on 14 July 2014.

From the statistics, cervical cancer can have devastating effects, with a very high human, social, and economic cost, affecting women in their prime. However, this disease should not be a death sentence as it takes years to develop from the precancerous stage to the severe stage. The use of low-tech and inexpensive screening tools could significantly reduce the burden of cervical cancer deaths.

Dr George Papanicolaou discovered more than 70 years ago that precancerous cells from the cervix could be identified before they turned lethal. He detected many asymptomatic cancer cases where some were in such an early stage that they were undetectable on a biopsy. The vaginal smear later became known as the Pap-smear. As Pap-smear screenings could detect cancer before it became symptomatic, they help early detection and enable prevention of the cancer from spreading and progressing to a severe stage. Early detection and treatment can prevent 75% of cancers from developing. In the UK, cervical cancer is the most common cancer in women under 35. The mortality rate from cervical cancer dropped significantly after the implementation of screening programmes. After the National Health Service Cervical Screening Programme started in the late 1980s, cervical cancer incidence rates decreased considerably. Cervical screening saves approximately 4,500 lives per year and prevents up to 3,900 cases of cervical cancer in England.

In western countries, the introduction of screening programs has successfully reduced incidence and mortality rates. Nevertheless, in less developed countries, women do not have access to effective screening activities. In Malaysia, the screening coverage is still poor, with the last statistics stating only a 2% screening coverage in 1992, 3.5% in 1995, and 6.2% in 1996. The reasons behind this poor screening coverage might be the uneven distribution of medical facilities and the shortages in human experts in relevant fields, i.e. pathologists and cytotechnologists.

The histological interpretation of tissue changes requires a medical background and specialized training. As a result, the availability of a histopathology service for cervical cancer screening activity depends directly on the availability of a pathologist. In addition to the problem of shortages in manpower, human error is an important issue which affects the screening activity. A possible way to handle these issues is to automate the screening and diagnosing process. With a high accuracy and user-friendly automatic cervical cancer screening system, the diagnosis process will be less dependent on the availability of pathologists or extensive medical facilities. Furthermore, automatic screening solves problems due to human error during the screening procedure such as fatigue or lack of diagnostic skill of some cytotechnologists. The development of an intelligent diagnostic system which complements the expertise of a human expert in the field, is becoming necessary towards reducing the incidence and mortality rate of cervical cancer in Malaysia.

By introducing NeuralPap, images captured from cervical slides can be processed using a computer and require shorter processing periods. Consequently, at higher screening frequency, more patients can be covered and this enables effective screening. This project aims to reduce the time consumed and labour needed in conducting a cervical cancer screening – and more importantly, to decrease the number of new cases and rate of mortality due to cervical cancer.

In order to achieve this aim, the two objectives of the NeutralPap project include:-

(a) To diagnose cervical cancer at an early stage with high accuracy of the diagnosis result.

(b) To reduce dependence on pathologists and cytotechnologists in screening activities through a user-friendly system.

By having this system, cervical cancer screening results will be available in a shorter period, thus making it possible to cover more patients in higher screening, which is important for making an effective screening activity possible.

MAKNA allocated RM 210,000.00 for this research in collaboration with the Faculty of Engineering, USM under research principal, Associate Professor Dr Nor Arshidi Mat Isa. 
(Article written by Prof Madya Dr. Nor Ashidi Bin Mat Isa, Faculty of Engineering, USM)