Incidence of Brain Atrophy and Decline in Mini-Mental State Examination Score After Whole-Brain Radiotherapy in Patients With Brain Metastases: A Prospective Study

Before the establishment of stereotactic radiosurgery (SRS), whole-brain radiotherapy (WBRT) was the golden standard of treatment for patients with brain metastases Currently, patients with single or oligometastases frequently are treated with SRS, whereas those with four or more metastases are considered to be indicated for WBRT; after SRS alone, the expected probability of tumor recurrence in the unirradiated areas is very high. Nevertheless, many patients with four or more metastases are treated by means of SRS alone without undergoing WBRT, especially in Japan . One of the major reasons for avoiding WBRT is the fear that WBRT may cause dementia, as well as brain atrophy. However, there are no data clearly indicating the incidence of such late adverse effects of cranial irradiation, and there are only retrospective studies suggesting the occurrence of these complications. Many patients reported previously were treated with surgery and radiation  therefore, it is unclear whether these complications are attributable solely to radiation therapy.

Brain atrophy and dementia may be related not only to surgery, but also to tumor status and chemotherapy . To properly evaluate the incidence of radiation-induced brain atrophy and dementia, we considered it necessary to carry out a prospective study to exclude as much as possible the influence of other factors. In this report, we present results of a prospective study of the Chubu Radiation Oncology Group, Japan (CROG-0301), that estimated the incidence of decrease in Mini-Mental State Examination (MMSE) scores and brain atrophy after WBRT in patients with brain metastases who did not undergo a neurosurgical operation or concurrent chemotherapy.

The apprehension that WBRT might cause brain atrophy and dementia seems to have grown gradually among medical oncologists and neurosurgeons. Several retrospective studies suggested it, but others reported maintenance of neurocognitive function in long-term survivors with glioma and other primary brain tumors after radiation therapy. Because retrospective studies cannot exclude the influence of other factors, such as surgery, chemotherapy, and disease progression, that can be associated with the development of brain atrophy and dementia, we conducted the present prospective study and attempted to eliminate as many of these factors as possible. As a result, we found that brain atrophy can develop in a proportion of patients, but decrease in MMSE scores was relatively infrequent. It was reported that other radiologic findings can develop after radiation to the brain, especially on MRI, but we used brain atrophy as an end point because it often is regarded as a late sequela of radiation linked to dementia. Another reason is that MRI was difficult to perform because of the long waiting time for booking in some institutions; however, brain atrophy could be evaluated easily by using CT.

We used the method of Nagata  to evaluate brain atrophy. It is a simple method that can be used in multi-institutional studies, but the widths of the lateral ventricles and sulci are influenced by the mass effect of the tumors. An increase in atrophy index was found in many patients at completion of radiation therapy. This was not caused by brain atrophy, but rather tumor shrinkage. Therefore, we used atrophy index at the completion of radiotherapy as a control to evaluate posttreatment brain atrophy in individual patients. Asai  reported the development of brain atrophy in 56% of patients undergoing radiation therapy. They used Nagata's method, as we did, but they defined development of brain atrophy as an atrophy index of 1.13 or higher. In our experience, measurement of CCR is not accurate enough to ensure that an atrophy index of 1.13 really represents brain atrophy. We believe an atrophy index of 1.3 is reasonable for visual recognition of brain atrophy on MRI and CT. In addition, all patients in the study of Asai  had undergone brain surgery. They reported that no brain atrophy was found after brain surgery alone, but it is not known whether surgery can be an additive factor in the development of brain atrophy when combined with radiation. In our study, excluding the influence of brain surgery, we found brain atrophy in up to 30% of patients at 6–18 months. However, about 40–50% of patients maintained an atrophy index of around 1 during these periods. We could not prove an association between the incidence of atrophy and patient age or use of the 10-Gy boost. Meanwhile, patients with a pretreatment CCR less than the median value had a greater incidence of increase in the atrophy index. This is in contrast to findings reported by Nieder showing that patients with preexisting atrophy had a greater risk of continuous deterioration. One reason for the finding in the present study may be that the index is likely to increase when the denominator (pretreatment CCR) is small.

The MMSE alone is considered to be an insensitive method to evaluate higher brain dysfunction, and it now seems clear that the combination of various neurologic tests is necessary to evaluate more subtle cognitive dysfunction. Therefore, the aim of the present study is to detect apparent dementia. In a prospective study comparing WBRT plus SRS and SRS alone, Aoyama evaluated changes in MMSE scores in a proportion of patients. They found that although there was no significant difference in change in MMSE scores after treatment between the two groups, the scores tended to decrease, especially after WBRT plus SRS. However, they did not clearly differentiate between disease progression–induced deterioration and treatment-related decrease. In the present study, mean MMSE score did not decrease on the whole, and proportions of patients with an MMSE score decrease of four points or more were only 11% at 6 months, 12% at 1 year, and 0% at 18 months. We excluded patients with intracranial progressive disease from further evaluation, but we did not exclude patients with systemic disease progression. As a consequence, about half the patients with an MMSE score decrease had systemic progressive disease and a decrease in performance status; thus, purely radiation-induced decrease appeared to be still less frequent. In the present study, only brain atrophy was evaluated by using MRI and CT, and there appeared to be no correlation between brain atrophy and MMSE score decrease. In additional investigations, we plan to evaluate MRI findings that may characterize patients with an MMSE score decrease.

Radiation dose per fraction may influence the occurrence of late morbidity for radiation therapy. It is well-known that central nervous system tissues have low α/β ratios and therefore are susceptible to greater doses per fraction. In WBRT for brain metastases, 10 fractions of 3 Gy commonly are used, but we did not use the 3-Gy/d dose in this study in the belief that a 2-Gy/d fraction is better than a 3-Gy fraction in terms of preventing late adverse effects in long-term survivors. Most previous studies reporting deterioration in neurocognitive function used 3-Gy or even higher doses per fraction  In addition, 10 fractions of 3 Gy given for prophylactic cranial irradiation in patients with small-cell lung cancer are considered to be more likely to produce neurotoxicity than 2- or 2.5-Gy/d fractions. In a Radiation Therapy Oncology Group study using 10 fractions of 3 Gy for brain metastases, 81%, 66%, and 57% of patients maintained an MMSE score higher than 23 at 6, 12, and 18 months, respectively. Although the biologically effective dose for 10 fractions of 3 Gy is less than that for 20 fractions of 2 Gy assuming an α/β ratio of 1–4 Gy, the incidence appeared greater than that observed in the present study. Thus, use of a 2-Gy/d fraction might have contributed to the favorable effects on neurocognitive function observed in the present study. Of course, we use 3-Gy fractions for palliative cases and patients with a short expected survival time, but we will continue to use the 2-Gy fraction for patients expected to survive longer than 6 months.

In summary, the present study shows that brain atrophy can develop after WBRT in a certain proportion of patients (up to 30%), but a decrease in MMSE scores was less frequent. Avoiding WBRT for the reason that it causes dementia appears to be a groundless idea in patients with metastatic brain tumors. The WBRT with or without stereotactic boosts should be a reasonable treatment for patients with multiple brain metastases.