Artikel
A multinomial logistic regression analysis on characteristics of hand and wrist tumours: A multi-ethnic study
Suche in Medline nach
Autoren
Veröffentlicht: | 6. Februar 2020 |
---|
Gliederung
Text
Objectives/Interrogation: Hand tumours are frequently encountered in clinical practice, however, large-scale epidemiological data is infrequently published. The existing literature is not homogeneous with differing accounts of the commonest tumours. Epidemiological data can provide diagnostic cues to guide the workup and management of hand tumours. Assessing significant independent demographic factors and tumour characteristics associated with incidence of hand tumours are essential in the planning and delivery of healthcare services.
Methods: A retrospective review of all patients between 2004 and 2015, who underwent excision of hand tumours in a tertiary hospital in Singapore, was conducted. The following data were collected: age, gender, ethnicity, histological diagnosis, malignancy and location of tumour. A multinomial logistic regression analysis was then performed.
Results and Conclusions: A total of 4476 tumours from 4226 patients were identified with a mean age of 51.3 (8-101 years), male to female ratio of 1:1.15. Majority of the patients were Chinese (75%), followed by Malays (9%), Indians (8%) and others (8%). The most common hand tumours excised were ganglions (43%), with the majority located in the wrist. The next most common were giant cell tumours (9%), which were most commonly found in the digits. The majority of soft tissue tumours were benign (97%), with only 3% malignant cases. Interestingly, a multinomial logistic regression analysis showed that the risk of malignant tumours originating from the skin was 22 times higher than from soft tissue, and 7.6 times higher than from bone for patients of the same age, ethnicity and sex. Also, the relative probability of tumours occurring at the metacarpals rather than fingers was about 1.32 times higher for males than for females and close to 2 times higher for patients with malignant tumours than those with benign tumours.
The overwhelming majority of soft tissue tumours of the hand and wrist are benign and this can guide workup as well as counselling of patients prior to operation. Malignant tumours, while the minority, have the potential for significant morbidity and mortality if not appropriately evaluated or treated. The application of a multinomial logistic regression analysis model observed that tissue origin from skin and tumour location over the metacarpals were significant predictors for malignant hand tumours.