The Utility of Radiotherapy in the Management of Haemoptysis Secondary to Aspergillomata and Structural Lung Diseases

Currently the best curative treatment of massive haemoptysis is surgical resection of the affected portion of lung. For patients who do not qualify for surgery, there are few alternative interventions, most of which only provide a temporary solution.

Surgical resection of the diseased areas, which offers the possibility of cure, is best performed as an elective procedure. It carries a not insignificant mortality, with reports varying between 1% and 50%. Unfortunately, many patients admitted to Tygerberg Academic Hospital, are not candidates for either elective or emergency surgical resection. The most common reasons for inoperability are: irresectable disease (i.e. damage to the lungs is bilateral and too extensive to allow resection, or the site of bleeding is not known); or severely reduced cardiopulmonary reserves secondary to extensive pre-existing lung damage, making lung resection surgery impossible due to excessively high mortality risk.

Repeat BAE (as a palliative measure in inoperable cases) is not always technically feasible and does not always lead to cessation of the bleeding. Additionally long term recurrence rates following BAE are variable estimated at between 18-42%, and carries a high mortality.

A significant number of patients with massive haemoptysis do not qualify for either surgery or BAE, thus rendering them without treatment option, save palliation with long-term opiates. Their fate is that of recurrent haemoptysis and a high associated mortality. Endobronchial occluding devices have been considered, however their high cost and required expertise prohibits their extensive use.

A novel treatment of this condition potentially is external beam radiotherapy (EBRT). A case report of five patients has documented EBRT in the setting of mycetoma where between 7 and 14Gy was used, with cessation of life threatening haemoptysis. However follow-up was limited to 6 months, and the effects of EBRT in the setting of post-tuberculous structural lung disease without mycetoma is not known. The potential mechanism of the achieved haemostasis was speculated as being the induction of radiation damage to the radiosensitive capillaries, with subsequent inflammatory response, in keeping with previous rat models of lung perfusion following high dose (30Gy) EBRT exposure.

There is therefore a paucity of data on the utility of EBRT in benign conditions such bronchiectasis and aspergillomata. Furthermore due to the different effects, particularly temporal, the radiation has on normal and benign conditions it is difficult to extrapolate from the malignant scenario to the benign. Laboratory studies have however demonstrated the effects of high dose radiation on tissue that may result in a reduced bleeding tendency.

This study aims to explore a novel therapeutic strategy for patients with life-threatening haemoptysis who are not candidates for definitive management (i.e. lung resection). This study also aims to establish if there is a measurable response to varying doses of chest radiation in cases of haemoptysis caused by benign conditions.

This study is a prospective randomised control-intervention study. Patients will be allocated to one of 2 arms (control and EBRT at 3.5Gy weekly for 5 fractions to a maximum of 17G). The primary outcome will be time to recurrent haemoptysis. Secondary outcomes include physical performance, lung function and the occurrence of EBRT complications. Patients will be followed up daily whilst in hospital and for one year post intervention.

Assessors and those performing the statistical analysis will be blinded to the group allocation. Statistical analysis will include univariate and multivariate analysis with appropriate parametric or non-parametric tests. Appropriate tests for categorical data (e.g. Chi-squared test) and continuous data (e.g. Kruskal-Wallis, and ANOVA) will be used. Logistic and linear regression modelling will be used for certain outcomes, and multivariate analysis will be performed using step-wise regression modelling and full modelling where appropriate.