Lower grade gliomas (LGG) are typically infiltrative and diffuse growing lesions, commonly involving eloquent regions (1–3). Although, slow progressing, they recur unavoidably and undergo malignant transformation (4). Despite better understanding of molecular patterns resulting in the new classification based on isocitrat dehydrogenase (IDH) mutation, surgery remains the main first line treatment (5, 6). A complete fluid-attenuation inversion recovery (FLAIR) or T2 based resection mostly allows for a longer progression free survival and may decrease rate of malignant transformation (7). Gross total resection or even supramaximal resection became an important goal for surgical treatment (1, 8–13). Nevertheless, the aggressive resection might result unintentionally to inferior quality of life (QoL) and compromise daily routines in both private and working spheres (14, 15). This holds true especially for eloquent lesions. Although, there are multicenter retrospective studies suggesting that a volumetric increase of extent of resection leads to an increased survival (5, 16). A deterioration of patients’ functional status apart from reduced QoL might lead to an exclusion from adjuvant treatment resulting in suboptimal outcome (17). Apart from counterbalancing of maximal safe resection and avoidance of neurological and cognitive deterioration, surgeons have to choose from a wide armamentarium of surgical tools various intraoperative imaging devices or mapping techniques at hand (10, 18, 19).

Currently, there are no randomized controlled trials (RCT) or controlled clinical trials (CCT) available on which to base clinical decision making (20). Informed consent is often based on surgeon’s individual experience (20). Incidence of neurological deficits in eloquent location like insular gliomas is often based on retrospective single center data and can thus be underestimated (21). The aim of our study was to evaluate the outcome of patients with eloquently located LGG based on prospective non-selected data from the Log-Glio registry.

Lower grade gliomas remain challenging neoplasms, since they affect typically younger patients and commonly infiltrate eloquent regions (2, 4). Tumor integration in neuronal networks may often limit extend of resection (23). There might even be a potential role of glioma’s molecular subtype influencing pathway disruption or displacement (24).

We have performed a detailed evaluation of eloquently located diffuse LGGs based on Log-Glio registry and focused on clinical outcome. Interestingly, despite functional intraoperative monitoring, we found relatively high number of patients with persistent new neurological deficit 3 months after surgery. Overall, almost half of the patients show new neurological deficits right after surgery. Both severe and mild neurological deficits show an improvement in half of these patients. Yet, around one third of patients permanently deteriorate both in neurological functions and in their daily life according to ECOG performance status. Predictive for permanent deficits was an impaired preoperative ECOG and a complete tumor resection in multivariable regression. Our data show that despite improved neuromonitoring and surgical techniques, risk for a permanent neurological deficit is high and improvement within the first 3 months is limited. This holds true, especially if a progressive resection aiming for a CTR is performed. The question arises; are mild neurological impairments justified if CTR is achieved? CTR has been shown to be an independent predictor for longer overall survival and even small tumor remnants could result in inferior survival (5, 16). From our perspective this question may only be answered for each individual patient. Our data may serve as a basis for patients’ informed consent before surgery to discuss potential risks and include them in decision making for surgical strategy. According to actual literature, the reported permanent deficits after glioma surgery range between 2-24% including both motor and speech deficits (5, 9, 12). Our series provides prospective multicenter data including only eloquent lesions. Most cited studies were monocentric and did not use clinical scores like NIHSS for detect deficits and hence may underestimate occurrence of postoperative deficits. This suggests that, the wide range of 2-24% of deficits might be more likely in the upper level when an unselected series is assessed. We found a prevalence of 38% at 3-month follow up. However, more than two-third of the patients improved within the first year of surgery. Especially, patients with a slight aphasia showed a good prognosis. In our series of eloquent tumors rate of intended CTR was relatively high with 48%, also intraoperative imaging was used in the majority of cases, suggesting a rather progressive approach compared to a slightly older series from the French glioma network with a CTR of only 12% including also non-eloquent lesions (25).

Surprisingly, even though only eloquent lesions were selected in our series in 20% of cases no use of IoM was reported which may bias our results. The prevalence of IoM in our series is higher than in the above cited contemporary study by Munkvold et al. (26) (76% vs. 58%). Given that only eloquent lesions were approached, from our point of view all patients should be operated using IoM. Even though, this finding is not supported by the results of the multivariable model. We hypothesize that, similar to awake surgery, with increasing risk of surgery, more likely IoM is used. Hence, protective effects may be leveled statistically.

Timing of surgery seems to play a role also in our data. There was a correlation of permanent new deficits in NIHSS and time to surgery >3month for primary surgeries. In the descriptive data no greater differences were found between patients operated early and patients watched and scanned except, that mostly accidental findings and tumors most likely being a WHO grade II lesion were watched for a longer time. This correlates with the current German management guidelines for glioma (27). Our data shows a slightly lower but relatively similar number of watch and scan than a current large Scandinavian series with 17% of patients (26). The authors have not provided the outcome data of these patients, so far. Based on the earlier findings from Jakola et al., being still the best evidence favoring an early surgical resection, one would expect lower occurrences of watch and scan (28). In our multivariate model ‘delayed’ surgery independently does not show a significant influence on new permanent deficits, but impaired preoperative ECOG did. From our point of view, it is important to avoid new deficits before surgery since it may mean that plasticity and redundancy of networks is already consumed by the tumor growth. Higher preoperative ECOG is a negative predictor factor for a good functional outcome, our data underlines the importance of an early resection before tumor progress results in a functional impairment.

Awake surgery is a gold standard for resection of eloquent located tumors adjacent to or in speech-eloquent cortical areas or fiber tracts (29, 30). However, awake surgery could be interpreted as a potential risk factor for permanent deficit according to our data. This result is limited by the relatively low number of awake surgeries in our series. Further, one might assume a selection bias for these patients since larger and more eloquently located tumors might more likely be operated awake. Compared to patients who had surgery asleep the proportion of preoperative deficits is relatively similar to patients operated awake, as is the rate of tumors involving the insula. On the other hand, preoperative ECOG is impaired more often in these patients reflecting the immense burden of impaired language function for patients. We interpret the potential association of awake surgery with permanent deficits not as a risk factor and we warn to draw the conclusion that asleep surgery is more protective. In direct comparison, current literature shows a superiority in extent of resection and occurrences of neurological deficits of awake to asleep surgery as reported by a current meta-analysis (21). Our data rather demonstrate the risk of operating language eloquent tumors, even when using the awake technique, which should be discussed with patients undergoing surgery in this area. Especially, patients with preoperative deficits seem to be at higher risk for permanent impairments as also found in other series (31). Our series, provides data 3 months after surgery which represents the typical reported time point for neurological deficits also after awake surgery (21, 32). Short term language deficits usually recover during this period (33). The proportion of long-term improvement of language deficits still present at 3 months was reported relatively low. In our series we found that especially slight aphasic deficits have a high probability of improvement within the first year (13, 34). Motor deficits had a more unfavorable outcome. The data of 9-12 months follow up was only available for 2/3^rd of patients with new deficits at first follow-up and for a better comparability to previously published data we calculated the regression models with the first follow up only. The remarkable occurrence of recovery in patients with aphasia may be biased by the detection using NIHSS. A false negative rate of 9% was reported by Grönberg et al. for stroke patients (35). Especially, a subtle anomic aphasia can remain undetected in NIHSS and may relevantly impair patients’ daily routines or return to work. Our results regarding ECOG also support this theory: Functional status based on ECOG showed a relatively stable course after surgery. Patients did not recover in overall function as they do on neurological deficits. Further patient reported outcome data like health related quality of life are needed to further address this issue.

In our study we did not find an influence of tumor biology as suggested by Young et al. (24). The authors showed an association of intra-tumoral function and pathway infiltration to molecular subtype of tumor. Whether it also influences surgical outcome has not been shown so far. In our series, both WHO grade and IDH mutations were also not influencing occurrence of permanent deficits. Most likely, effects are subtle and larger series will be needed to further address this question.

Intraoperative imaging like intraoperative MRI and intraoperative ultrasound were shown to increase extent of resection, and are widely used by European neurosurgeons as also in the participating centers in our study (10). Yet, in this eloquent series proportions of CTR are lower than previously published for intraoperative imaging (36–38) since resection is obviously limited by function even in cases in which a CTR was deemed feasible preoperatively. Intraoperative imaging like iMRI was applied more often in patients with an intended CTR. Yet, while CTR highly correlated with permanent deficits and also was an independent predictor for them, intraoperative imaging did not correlate with new permanent neurological deficits in our series. One of the reasons could be that in all cases iMRI was used, it was combined with IoM, for iUS is was slightly lower with 82%. Further, an intraoperative visualization of individual anatomic structures after relevant brain shift may also increase safety. However, no evidence for this hypothesis can be found in the current data.

Interestingly, recurrent surgery showed no significantly higher risk of deficits. One explanation may be the relatively low number of cases in our series. Regarding recurrent surgery, the concept of multiply staged approaches as proposed by Duffau et al. (34) may be a relevant strategy. Cortical reorganization in relation to function by neuroplasticity may increase safety in a second surgical approach. A detailed preoperative functional imaging by functional or resting state MRI, integration of connectomics and/or non-invasive cortical mapping may foster thorough preoperative evaluation in these patients and may increase safety of resection during surgery (39–42). Further studies are needed to evaluate influence of this preoperative data on surgical outcome. Another future perspective is to study patient reported outcome measures as health-related quality of life or supportive care need to better understand how fine cognitive difficulties or motor deficits influence patients’ daily life or occupational situation.

Patients harboring eloquently located LGG are highly vulnerable for permanent deficits. Almost one third of patients have a permanent reduction of their functional status based on ECOG. Risk of an extended resection has to be balanced with the respective oncological benefit. Especially, patients with impaired pre-operative status are at risk for new permanent deficits. There is a relevant improvement of neurological symptoms in the first year after surgery, especially for patients with slight aphasia.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

Ethical approval was received by the ethic committee of the University of Ulm (Ethikkommission Ulm, No. 201/15). The patients/participants provided their written informed consent to participate in this study.

Conceptualization, JC, CW, AP, GN, DR. Methodology, JC, AP. Formal analysis, JC. Investigation, JC, JO, MU, CR, KK, MN, M-TF, RG, ML. Resources, CW, VR, MC, OG, PV, MT, DR. Data curation, JC. Writing—original draft preparation, AP, JC. Writing—review and editing, all contributing authors. All authors have read and agreed to the published version of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.