Glioblastoma remains an incurable tumor that requires aggressive surgical and adjuvant treatments in order to prolong local disease control and OS 4,28,29. Maximal safe resection is a gold standard while benefits of proposed supramarginal resections are under investigation with promising results30.
However, it is important to shed a light on white matter boundaries and eloquent areas: the more extensive the resection, the higher is the risk of postoperative morbidities11. Considering a meta-analysis by De Witt Hamer et 31, the risk for late post-operative neurological deficit ranges from 3.4–8.4% of treated patients, while neurological deficits at discharge can reach nearly 30% of patients according to different series28,32,33. Such risks may increase in case of supramarginal resection 34 and it is under debate if such risk worth the potential benefits 30,35–37: in a work by Aabedi et al16 neurological integrity was found to be a key factor for survival.
Although it is commonly known that post-operative KPS and deficits affect the neuro-oncological outcome of patients with gliomas, few studies in literature investigated how these deficits impact on access to adjuvant therapies 38–40. However, it is a common finding that short delays of chemoradiation therapy do not affect overall outcomes; on the contrary, several reports show that both early chemo- or radiation therapy and long delays can negatively affect neuro-oncological outcomes 13,41–45.
In order to better understand how neurological deficits impact on prognosis and access to adjuvant therapies for GBM patients, we performed a stratification of pre and postoperative deficits on a homogeneous cohort. Our case series showed that 81.4% of patients presented with a preoperative deficit; however, after surgery 83% of them showed stability or improvement of deficit. Overall, nearly 20% of our patients were not able to access adjuvant treatments.
The surgical strategy was independent from the presentation with a deficit (p = 0.490) and from the kind of deficit at presentation (p = 0.907). The presence of a deficit at diagnosis showed to slightly reduce PFS (p = 0.04) and OS (p = 0.08): furthermore, both presentation with hemiparesis and aphasia at diagnosis related to a reduced PFS (p = 0.03) and OS (p = 0.04) while other minor deficits as well as hemianopia showed curves similar to patients without symptoms at presentation. This discrepancy was more and more evident when considering curves stratified by deficit at discharge: the presence of a neurological deficit at discharge was found itself to negatively impact on OS (p = 0.01) and showed a shorter, yet not significantly, median PFS of 6.27 months Vs 9.77 months (p = 0.2). PFS was not associated with the improvement or worsening of preoperative deficits (p = 0.1), while OS was longer for patients with improvement of neurological deficits at discharge (p = 0.02). This discrepancy could be explained as follows: early progression in patient with a deficit depends exclusively on surgical resection and on the natural history of the disease. On the other hand, in a homogenous cohort, OS depends both on surgical resection and on access to adjuvant treatment, which are reserved only for patients able to get to the hospital while the others are considered too fragile to undergo chemo- or radiation-therapy.
Despite these data, the chance and timing to access adjuvant therapies was the same for patients with or without deficit at presentation (p = 0.933), both for chemotherapy (p = 0.381) and for RT (p = 0.848). Nevertheless, impairment of clinical condition at discharge significantly reduced the chance to get to adjuvant therapies for patients (p = 0.004).
In our series, clinical impairment at discharge and KPS < 70 resulted to be significantly associated with the lack of access to neuro-oncological cares (p < 0.001) and consequently an important prognostic factor for OS: median OS for patients with KPS ≥ 70 was 19.80 months and almost halved for patients with KPS < 70(8.80 months; p < 0.001) 46.
Moreover, some patients due to their acquired neurological deficit may be not able to be assisted at home during adjuvant treatments: such situation may explain why patients suffering from aphasia or hemiparesis had a reduced access to adjuvant treatments.
In our cohort the worst prognostic factor was hemiparesis at discharge, highly decreasing the chance to get to adjuvant therapies and survival, followed by aphasia. From our results, a motor weakness severely impacts on OS and PFS more than other neurological deficits (both p < 0.0001); this influence may be explained with the severe difficulties that these patients have to deal with in daily life. Second of all, aphasia showed to be associated with the second worst PFS and OS (p < 0.0001) probably due to impact that such deficit can have on quality of life and on compliance to treatments. Our findings confirmed the findings by Quinones-Hinojosa 47. Although in their work they did not analyze the access to adjuvant therapies, they found that a new neurological deficit at discharge reduced of at least 6 month the life expectancy 8. Interestingly, patients with homonymous hemianopia experienced PFS and OS like patients with no deficit at discharge. Such finding may be explained by the fact that homonymous hemianopia may not significantly influence the quality of life and independency at discharge, which results in a comparable ability to get to adjuvant treatments and comparable KPS. Moreover, patients harboring right temporal or occipital tumors may be treated more aggressively with supramarginal resections consisting mostly in lobectomies as suggested by Schneider et 38, at low risk of neurological deficits affecting normal daily activities 13. In this view, homonymous hemianopia might be an acceptable acquired neurological deficit for patients affected by a GBM; especially after a supramarginal resection. In terms of resection, patients who received a GTR had the longest median OS of 18.40 Vs 17.17 months for STR, 10.70 months for PR and 4.33 months for patients who underwent stereotactic biopsy. (p < 0.05). Eleven (50%) patients who underwent biopsy were not able to undergo to adjuvant therapies, the other 50% started RT with chemotherapy with TMZ but only 4 patients ended the first adjuvant chemotherapy cycle after RT. Despite the small sample size of the biopsy cohort, we found a selection bias which could lead itself to a worse prognosis: almost 82% of patients selected for biopsy presented with tumor involving more than 2 lobes and the remaining 27% had tumor extension to the midline. Furthermore, almost 90% of patients selected for biopsy, showed a deficit at presentation. As described in the work of Pasqualetti et al 48 this selection bias of patients already compromised at diagnosis, generates itself a survival bias; and in our study a bias for access to adjuvant therapies.
However, in a landscape where biopsy is related with significantly lower OS and PFS 28,49, and significantly reduced chances both in terms of access to adjuvant treatments and quality of life, it is fundamental to ask ourselves: is there really a need of a diagnosis for the most defeated patients?
Limits of the study
Our study has some limitations. First of all, it is a retrospective study that may be influenced by revision of clinical records. Moreover, since neuropsychological evaluation of patients with gliomas has started in 2019, cognitive deficits were not recorded and may have influenced access to adjuvant therapies more than a deficit itself, especially in case of elderly patients. In fact, patients with brain tumors are susceptible of cognitive deficit and may not recover after surgery. Moreover, for older patients, the cognitive status can negatively impact on access to adjuvant treatments14.