| Purposes:Intraoperative direct electrical stimulation(DES)combined with neuromonitoring(IONM),the gold standard approach for the mapping of eloquent cortices and tracts during the resection of gliomas,paralleled with awake craniotomy(AC)and intraoperative functional tasks,has become the common standard surgery strategy of gliomas involving eloquent areas.With the development of f MRI,neuronavigation and intraoperative MRI(i MRI),maximal safe resection of gliomas involving eloquent areas becomes possible under general anesthesia(GA).Large amount of studies had proved good outcomes of patients after surgery under GA,while the choice of AC/GA is still controversial in the treatment of gliomas involving eloquent areas(GIEAs).This study aims to compare clinical and tumor features of gliomas involving different functional areas(Gi DFA)in our center.Then to compare whether surgery under GA or AC will cause different outcomes of Gi DFA,we summarize the outcomes and survival of Gi DFA after resection assisted by different modal techniques under GA,compare them with those of Meta-analysis results of previous GA/AC studies.Multivariate analysis is used to build prediction equations of incidences of postoperative temporary/permanent neurological(motor,language,visual)deficits,finally the GA/AC choice model is set up for gliomas of eloquent areas based on the previous comparison results and prediction equations so as to provide guidance to surgeon and patients preoperatively.Taking glioblastoma invading corpus callosum(cc GBM)as case samples,we preliminarily explore whether multimodal surgery under GA can protect advanced neurological function.Methods:Cases of gliomas involving eloquent areas from January 2009 to December2020 in our department were collected and divided into Group A(motor areas),B(language areas)and C(visual areas).Their clinical and tumor features were compared.Groups A,B,C were further divided into Groups A1,B1 and C1(using neuronavigation only),A2,B2 and C2(using neuronavigation,DES and IONM),A3,B3 and C3(using neuronavigation and i MRI),A4,B4 and C4(multimodal group,using neuronavigation,i MRI,DES and IONM)respectively.The postoperative outcomes of all the groups were compared with each other,the outcomes included EOR,rate of gross total resection(GTR),surgery duration,length of hospital stay,incidence of surgery related complications,incidence of seizure,neurological function related parameters(KPS,muscle strength,Aphasia Quotient(AQ),visual impairment at 1 week/3months/6 months postoperatively),incidences of temporary/permanent neurological deficits,progression-free survival(PFS)and overall survival(OS).If the i MRI was scanned more than once and the residual tumor was further removed,the EORs were compared between the first scan and the final scan.Patients of GIEAs were divided into 2 groups,Group 1:temporary/permanent neurological deficits occurred,Group 2:temporary/permanent neurological deficits did not occur.Gliomas of motor areas were divided into A1 and A2,gliomas of language areas were divided into B1 and B2.The time points of temporary and permanent neurological deficits were defined as 3 months and 6 months postoperatively.The clinical and tumor features were compared between A1 and A2,B1 and B2 to identify significant variables by using univariate analysis.Then significant variables were included in multivariate analysis of Logistic regression,finally we aimed to identify significant variables associated with the occurrence of temporary/permanent neurological deficits and build the prediction equations.Probabilities of temporary/permanent neurological deficits occurrence were calculated according to prediction equations.We reviewed previous studies of GA and AC of GIEAs,performed Meta-analysis on their incidence of temporary/permanent neurological deficits,postoperative seizure and complications.The PFS and OS of these studies were also summarized.Our results of GA or AC were compared with these results(95%confidence interval[CI])of Meta-analysis.According to the comparison results and the prediction equations,the predictive probability threshold was determined and choice model of GA or AC was established for glioma involving motor or language areas.It helps to decide GA or AC for different patients.Meanwhile receiver operating characteristic(ROC)curves analysis can determine the best predictive probability threshold and assess the accuracy of choice model.The cc GBM patients of GA were included between January 2016 and December 2020,they were divided into group 1(using neuronavigation only)and group 2(multimodal group).Besides outcomes referred above,the Mo CA scores were compared between groups,the preoperative and postoperative Mo CA scores of each group were also compared to explore the protective effect of resection assisted by multimodal techniques on advanced brain function.Results:One thousand one hundred and forty-two gliomas of motor areas,564 gliomas of language areas and 279 gliomas of visual areas were included.Compared with the other two groups,patients of gliomas involving visual areas had older median age(52 years),larger median tumor volume(52.65cm~3),more GBM(63.8%)and less LGG of WHO grade 2(12.9%).Headache,nausea were more common among these patients.Preoperative KPS scores were:motor area>language area>visual area.The median EOR and GTR rates in A3 and A4 groups were higher than those in A1 and A2 groups,while the incidences of temporary/permanent motor deficit in A4 were lower than those in A1(17.3%vs 28.9%,P=0.001;6.9%vs 23.6%,P<0.001).The neurological functions of 6 months postoperatively were still significantly improved compared with those of 3 months,so the time point of permanent neurological deficits was defined as 6 months postoperatively.For gliomas of all grades,the median PFS and OS in A4 group were significantly longer than those in A1 group(P<0.05).For high grade glioma(HGG),the median PFS and OS in A3 group were also longer than those in A1 group.The median EOR and GTR rates in B3 and B4 groups were higher than those in B1 and B2 groups.The incidences of temporary/permanent language deficits in B3,B4 groups were lower than those in B1,B2 groups.For HGG,the median PFS and OS in B3 and B4 groups were longer than those in B1 and B2 groups(P<0.05).C3 group had higher median EOR and GTR rate than C1 and C2 groups,while C1-C4 groups had similar postoperative incidences of visual deficit,KPS.For HGG,the median PFS and OS in C3 and C4 groups were longer than those in C1 and C2 groups(P<0.05).Due to multiple use of i MRI,the final median EORs of gliomas involving motor,language and visual areas increased by 9.49%,10.53%and 11.28%respectively,they were all significantly higher than the EORs of first i MRI scan(P<0.05).Prediction equations of temporary/permanent motor deficits incidences were ln(P/1-P)=-1.625+0.015X1(age)-0.204X2(nearest distance between tumor and motor area)-0.696X3(involving the premotor or supplementary motor area[PMA/SMA])-0.449X4(DES/IONM),ln(P/1-P)=-1.898-0.289X1(nearest distance between tumor and motor area)-0.51X2(i MRI)-1.012X3(DES/IONM)respectively.Prediction equations of temporary/permanent language deficits incidences were ln(P/1-P)=-3.59-0.175X1(nearest distance between tumor and language area)+0.039X2(preoperative AQ)-0.885X3(i MRI),ln(P/1-P)=-3.428-0.202X1(nearest distance between tumor and language area)+1.241X2(involving PMA/SMA)+0.032X3(preoperative AQ)-1.065X4(i MRI)respectively.The GTR rate of A4(75.1%)group were higher than the Meta-analysis results of previous GA(53.8%[95%CI:43.6-64.1%])and AC(49.7%[42.1-57.3%])studies.Its temporary motor deficit incidences were 17.3%,which was lower than the 95%CI of Meta-analysis result of GA studies(28.3%[20.2-36.3%])but was at the lower limit of the result of AC(23.4%[17.3-29.5%]).Permanent motor deficit incidence of A4 group was 5.5%which was within the95%CI of GA(7.6%[5.3-9.9%])and AC(7.6%[5.1-10.2%])studies.The GTR rates of B3(72.5%)and B4(72.1%)groups were higher than the Meta-analysis results of previous GA(44.1%[26.4-61.9%])and AC(46.2%[37.3-55.1%])studies,their temporary language deficit incidences(17.2%,20.0%)were both similar with Meta-analysis results of GA(15.4%[9.2-21.7%])and AC(21.8%[16.1-27.5%]).B3(13.7%)and B4(13.1%)groups had the higher incidences of permanent language deficit than the Meta-analysis result of AC(7.7%[4.1-11.3%]),but they were not different with that of GA(12.4%[5.8-19.1%])studies.Compared with the results of previous GA and AC studies,the PFS and OS of LGG and HGG in groups A3,A4,B3 and B4 were all at a higher level.According to the comparison results above,multimodal techniques-assisted surgery under GA can completely replace AC for gliomas involving motor areas.The best probability threshold of prediction equation of temporary motor deficit was 23.3%,Sen was 0.69,Spe was0.67 and AUC was 0.71(95%C I:0.67-0.75).The best probability threshold of prediction equation of permanent motor deficit was 12.5%,Sen was 0.79,Spe was 0.67 and AUC was0.78(0.74-0.82).The best probability threshold of prediction equation of temporary language deficit was 22.0%,Sen was 0.73,Spe was 0.58 and AUC was 0.69(0.64-0.74).The best probability threshold of prediction equation of permanent language deficit was 17.1%,Sen was 0.71,Spe was 0.65 and AUC was 0.72(0.66-0.78).The reference probability threshold for gliomas involving language areas was 11.3%.If the neuronavigation and i MRI are certain to be used and the predictive incidence is higher than 11.3%,AC is recommended.The AUC will be 0.75(0.67-0.83),Sen will be 0.84,Spe will be 0.51.The multimodal group had higher median EOR(P=0.036)and GTR rate(60.7%VS33.3%,P=0.032)than group 1(using neuronavigation only)for cc GBM patients.There was no significant difference in MOCA score and KPS between two groups at different time points postoperatively.The multimodal group had longer median PFS(P=0.023)and OS(P=0.044).Higher EOR,radiotherapy and longer chemotherapy cycles were significantly associated with the survival of cc GBM.The EOR threshold 92%at least can prolong PFS and OS significantly.Conclusions:For patients of gliomas involving motor,language and visual areas,multimodal techniques-assisted surgery can achieve the better outcomes under GA.The time point of permanent neurological deficits was defined as 6 months postoperatively.The multiple use of i MRI can increase the median EOR by 10%.Multimodal techniques-assisted surgery under GA is enough for the resection of gliomas involving motor areas,it enable patients achieve similar outcomes with AC and a higher EOR and longer survival.If the predictive probability of permanent language deficit≤11.3%,multimodal techniques-assisted surgery can be performed under GA for patients with glioma involving language areas.Otherwise,it is recommended to perform AC combined with multimodal techniques for these patients to preserve their language more effectively.The established choice model can help surgeon and patients make appropriate and individualized choice of GA or AC for resection of gliomas involving eloquent areas according to different clinical and tumor features.This model has the moderate accuracy in predicting the incidence of temporary/permanent motor or language deficits.Advanced neurological function can also be preserved by multimodal surgery under GA,but it needs further prospective study to be proved. |
PDF Full Text Request