The value of Apparent Diffusion Coefficient as a prognostic biomarker in brain metastases treated with radiosurgery
We investigated the value of Apparent Diffusion Coefficient (ADC), derived from MRI, as a prognostic biomarker for local response and overall survival in 202 patients with brain metastases from non-small cell lung cancer (NSCLC) treated with stereotactic radiosurgery. MRI scans was performed before treatment and at two follow-ups.
Our study did not find pre-treatment ADC of brain metastases from NSCLC to be a significant predictor for neither local control nor overall survival.
The incidence of brain metastases is increasing, occurring in 9-17% of all cancer patients. Amongst adults, it is the most common intracranial tumour, with lung cancer being one of the most frequent malignancies spreading to the brain. The increase in brain metastases is 3 of 20 partly due to better diagnostic methods, as well as improved survival with cancer. However, the overall prognosis for cancer patients with brain metastases remains poor, thus better diagnostics and treatment is crucial. Primary treatment options for brain metastases are surgery, stereotactic radiosurgery (SRS) and whole brain radiotherapy (WBRT). Conventionally the treatment response in radiation therapy relates to change in tumour size as diameter. This, however, does not occur until several months after treatment, if it happens at all. Radiation related changes can cause pseudo progression of the tumour, such as central necrosis, which can cause the tumour to increase in size or fail to shrink. These changes could easily be interpreted as treatment failure even though this is not the case. A potential biomarker for early detection of treatment response is Apparent Diffusion Coefficient (ADC). ADC is a measure of the water diffusivity within a tissue and is calculated using diffusion weighted magnetic resonance imaging (DW-MRI). Studies performed in different tumour types have shown that there is an inverse correlation between ADC and the level of cellularity within a tissue. ADC has shown potential of earlier prediction of response to treatment in various cancers, including brain metastases. ADC has been shown to detect a response as early as 7-9 days into radiotherapy and is also able to differentiate between pseudo progression and actual treatment failure. Furthermore, studies performed in cerebral metastases and other tumours have shown that tumours with a low pre-treatment ADC tend to respond more favourably to SRS and chemotherapy than tumours presenting with high pre-treatment ADC. Other studies have shown that an increase in ADC was associated with an improved response rate after treatment. ADC lacks investigation in regards to survival. However a study found that low ADC was predictive of poor brain metastasis-specific progressive-free survival in breast cancer patients. Another study has also found ADC to be a potential prognostic factor for overall survival in cervical cancer, where both pre-treatment ADC and change in ADC from pre-treatment to approximately 3 months posttreatment has been shown to be predictive for survival. Earlier prediction of response could allow for more individualized treatment such as boost treatment (additional radiation dose to ADC defined regions) or change in overall treatment strategy, thereby avoiding tumour regrowth and proliferation of the tumour cells, potentially bettering the treatment outcome for the patient. Better and faster insight to tumour response can potentially enable better local control and reduce treatment toxicity. The aim of this study was to investigate the value of ADC as a prognostic imaging biomarker for local response and overall survival in patients with brain metastases from non-small cell lung cancer (NSCLC). Specifically, the added value of the ADC in conjunction with standard clinical data was investigated.
Description of the cohort
In this retrospective cohort study, we investigated patients with NSCLC and brain metastases, treated with SRS at 20 Gy in one fraction in the period January 2012 to December 2019 in the Region of Southern Denmark. The lesions which received this treatment were considered target lesions. Brain metastases within the same patient, if any, which received non- or other dose of SRS was considered non-target lesions. We did not assess response or characteristics of the non-target lesion, but they were included in the evaluation of the overall tumour burden in the patient. All patients received corticosteroids as a part of treatment on the day of SRS. Depending on symptoms some patients also received corticosteroids prior to and after treatment. Exclusion criteria were any prior treatment to the target lesions, including WBRT. All patients meeting the eligibility criteria were included, resulting in 202 patients with a total of 288 brain metastases.
Data and biological material
Data was collected at baseline and at two follow-ups. Baseline was defined as the time of the simulation MRI scan, performed as a part of treatment planning alongside a simulation CT scan. In a few cases no MRI simulations scans were performed, and the diagnostic MRI scan was used in treatment planning instead. In these cases, the diagnostic scan was used as baseline. All patients were treated at Odense University hospital, but follow-up scans could be performed at the patient's local hospital. For all patients, demographic data, clinical data, and tumour characteristics were collected. Demographic data and clinical baseline data were collected from medical records. Tumour characteristics were collected from the MRI scans, the radiology descriptions, and medical records. The clinical baseline data consists of 'the time difference between diagnosis of NSCLC and diagnosis of brain metastases' (TBD), performance score (PS) as defined by The World Health Organisation (WHO), activity of extracranial disease, presence of extracranial metastases, smoking, and pathology of lung cancer. Presence of extracranial metastases was evaluated as by the 8th edition of the TNM classification for non-small-cell lung cancer. Activity of primary disease was defined as treatment to primary lung cancer and/ or extracranial metastasis during control period and/or within 3 months before baseline. Demographic data consist of sex, age, and body mass index (BMI). Tumour characteristics consist of gross tumour volume (GTV), pre-treatment ADC calculated from both GTV and viable tumour volume (VTV), which is a DW MRI defined tumour sub volume with low diffusivity (indicating high cellularity), total number of brain metastases (target and non target lesions), and widest diameter.