Glioblastomamultiforme (GBM) is the most common and deadliest primary brain tumor.14The incidence of the disease is 5 to 10 cases per 100,000 persons per year, and more than 14,000 new cases of GBM are diagnosed in the United States each year.7Although new treatment paradigms have had a significant impact on the outcomes of many other cancer types, the treatment of patients with GBM remains a challenge. Patients survive for an average of 14 months from diagnosis despite maximal treatment with surgery, chemotherapy, and radiation therapy,23和duration of survival has not changed significantly in decades.
The role of surgery in the management of GBM has been extensively investigated. Initial studies traditionally classified extent of resection (EOR) within the arbitrary categories of gross-total, subtotal, and partial resection. Our previous study on 416 patients with newly diagnosed and recurrent GBM shifted that paradigm of the 3 arbitrary categories into the concept of maximal, safely achievable volumetric resection, demonstrating an incremental increase in the duration of survival with each unit increase in the EOR, starting from 89%,15with the strongest effect of resection on survival being achieved at the 98% threshold. Years later, these results were supported by investigators from the University of California, San Francisco (UCSF) in a retrospective review of 500 consecutive patients with newly diagnosed GBM, with minor differences in the EOR threshold, probably reflecting the distribution of EOR between the 2 study samples.20The UCSF data showed a significant survival advantage starting at 78% EOR instead of the 89% reported in our series, and an EOR ≥ 95% had the greatest impact on overall survival instead of the ≥ 98% from our series. The median survival time among the 316 UCSF patients in the ≥ 95% EOR category was 14.5 months.20Our group recently revisited data from 721 adult patients with GBM from 1993 to 2012 and provided further evidence supporting a maximum safe resection approach for GBM.19
Glioblastoma multiforme is regularly described as an invasive tumor.26The motility of GBM cells has been demonstrated in vitro in cell cultures,2,12in postmortem studies,3,4,27and in stereotactic biopsy specimens,10,13as well as in other studies.5,17,18,24The area outside the contrast-enhancement region on a T1-weighted MR image is usually infiltrated by tumor cells.26Fluid-attenuated inversion recovery (FLAIR) images are thought to represent these invasive cells,8,13as well as cerebral edema, demyelination, and/or surgery-related injury.9,28In a general review of GBM, Wilson points to the inadequacy of the current treatment paradigms for the patient presenting with GBM and links the probability of recurrence at any site to the number of cells left behind at or around the initial GBM site,26supporting his theory by the fact that 80%–90% of patients who succumb to GBM die of local tumor recurrence,4,6,11,25unlike those with other malignant tumors who die of metastasis.
如果增加contrast-enhanc的切除ing portion of a GBM leads to increased survival, based on the above 2 studies and other supportive literature, does 100% resection of the contrast-enhancing tumor confer a significant survival advantage relative to that obtained following the less expansive, though still extensive, resection? And would the additional removal of infiltrating tumor cells, as identified by FLAIR abnormality on T2-weighted MR images, have a significant impact? To answer these 2 questions, we identified patients undergoing an EOR ≥ 78% at our institution. We first compared the survival of patients with 78% to < 100% EOR with that of patients undergoing 100% EOR of the enhancing mass (with or without additional resection of the FLAIR abnormality). An EOR of 100% of the enhancing lesion is hereafter referred to as complete resection. The 78% cutoff was chosen because it was the lower of the 2 thresholds most significantly associated with survival in the prior MD Anderson and UCSF studies.20In a secondary analysis, we reviewed the effect of the additional resection of the FLAIR abnormality region on survival within the group of patients undergoing complete resection of the enhancing lesion. The incidence of postoperative complications—overall and neurological—was also reviewed.
开云体育世界杯赔率
Patients and Treatment Characteristics
The study was conducted under the auspices of an institutional review board-approved protocol. In a search of the Brain and Spine Center Database at the University of Texas MD Anderson Cancer Center (MD Anderson), we identified all consecutive patients with GBM who had undergone tumor resection between June 1993 and December 2012. All cases had a diagnosis of GBM according to the Ringertz classification, Grade IV astrocytoma according to the WHO classification, and Grade 4 astrocytoma according to the St. Anne-Mayo classification.15We excluded patients having more than 1 GBM, those aged 80 years or more at the time of surgery, and those with < 78% contrast-enhancing tumor resection. The final sample consisted of 1229 patients with GBM.
Demographic, clinical, and imaging data were obtained from the prospectively collected Brain and Spine Center Database. Each patient’s age, sex, and Karnofsky Performance Scale (KPS) score at the time of surgery was obtained. Previous neurosurgeries were recorded, including those undergone at another center, including cytoreductive surgery or a biopsy procedure with or without adjuvant chemotherapy or radiation therapy, which indicated that the patient had residual or recurrent tumor at the time of presentation at our institution. Such patients are referred to hereafter as the previously treated group. The MD Anderson Tumor Registry was used to identify each patient’s vital status at the time of data analysis. The tumor registry staff ascertain the vital status of all patients seen at our institution through sources such as the National Death Index and from letters or phone calls to patients and their families.
Imaging Studies
Preoperative and postoperative MR images (typically obtained within 48 hours after surgery) are routinely reviewed at our institution in a prospective fashion. Several tumor imaging characteristics are identified and entered into the Brain and Spine Center Database. Tumor location in regard to proximity to eloquent brain is characterized by functional grade, as described by Sawaya et al.21The presence of mass effect, surrounding edema, and contrast-enhancement of the tumor mass are also assessed. Preoperative and postoperative tumor volumes are quantified. Tumor contrast-enhanced volume is defined as the area of increased signal intensity on contrast-enhancing T1-weighted MR images. The FLAIR abnormality volume is defined as the area of FLAIR hyperintensity signal seen on T2-weighted MR images beyond the contrast-enhanced T1-weighted images. The Vitrea software version 2 is currently used for these studies (the MedVision 1.41 computer software program from Technologies Inc. was used for older cases). This software allows calculation of the tumor area as outlined on selected axial or coronal images and then the estimation of tumor volume based on the known thickness of the slice.
Statistical Analyses
本研究的主要目标是establish whether complete resection of the contrast-enhancing lesion (100%), as determined on the basis of objectively quantified preoperative and postoperative volume estimates, has prognostic value with regard to patient survival time after surgery, compared with resection < 100%. The secondary objectives were to examine whether additional resection of the region corresponding to the FLAIR abnormality beyond the contrast-enhancing area has a significant effect on survival time overall or within defined patient subgroups and to assess the incidence of overall and neurological complications associated with different EOR scenarios.
To achieve these objectives, frequencies and descriptive statistics of the variables under study were obtained. A recursive partitioning analysis was performed to identify a cutoff for continuous variables, if one existed. The cumulative survival duration measured from the time of surgery at our institution was computed using the Kaplan-Meier method. Survival curves for the various subgroups were compared using the log-rank test. The Cox proportional hazards model was used to identify factors associated with survival at the univariate and multivariate levels. Crude and adjusted hazard ratios and their 95% confidence intervals were calculated. All p values ranging from 0.0001 to 0.1 for categorized continuous variables were adjusted to achieve an approximate false-positive rate of 10%, based on the paper by Altman et al.1All tests were 2-tailed. A p value ≤ 0.05 was considered significant. Statistical analyses were performed using the Statistical Package for the Social Sciences 21.0 (SPSS Inc.). The recursive partitioning analysis was performed using the R software (version 3.1.0).
Results
Patient and Tumor Characteristics
Table 1summarizes the demographic, clinical, and tumor characteristics of the study patients. The median age was 55.7 years (range 4.6–80.0 years). Most patients were male (62%). The median KPS score was 90 (range 10–100), and 92% of patients had a score of at least 70. Most patients (91%) had symptoms before surgery. Previously untreated patients accounted for 61% of the group. Based on preoperative MR images, 14% of tumors were presumably located in noneloquent brain (Grade 1), 48% were located in near-eloquent brain (Grade 2), and 38% were within eloquent brain area (Grade 3).21Necrosis was documented in 78% of tumors; a cystic component was found in 10% of tumors. The median preoperative contrast-enhancing tumor volume was 31.0 cm3(range 0.3–186.3 cm3).
Characteristics of 1229 patients with GBM*
Characteristic | Total No. | 100% Resection of T1 Contrast-Enhancing Vol | 78% to <100% Resection of T1 Contrast-Enhancing Vol | p Value |
---|---|---|---|---|
No. of patients | 1229 | 876 | 353 | |
Median age at surgery in yrs (min, max) | 55.7 (4.6, 80.0) | 55.5 (4.6, 80.0) | 56.7 (6.0, 79.5) | 0.2 |
Sex (%) | ||||
Male | 758 (62) | 539 (62) | 219 (62) | |
Female | 471 (38) | 337 (38) | 134 (38) | 0.87 |
Median preop KPS score (min, max) | 90 (10, 100) | 90 (20, 100) | 80 (10, 100) | <0.001 |
Preop KPS score (%) | ||||
≥70 | 1134 (92) | 819 (93) | 315 (89) | |
<70 | 95 (8) | 57 (7) | 38 (11) | 0.01 |
Tumor functional grade (%)† | ||||
1 | 166 (14) | 151 (17) | 15 (4) | |
2 | 589 (48) | 409 (47) | 180 (51) | |
3 | 469 (38) | 313 (36) | 156 (44) | <0.001 |
Symptoms before surgery (%) | ||||
No | 105 (9) | 85 (10) | 20日(6) | |
Yes | 1124 (91) | 791 (90) | 333 (94) | 0.02 |
Previously treated (%) | ||||
No | 752 (61) | 539 (62) | 213 (60) | |
Yes | 477 (39) | 337 (38) | 140 (40) | 0.70 |
Necrosis on imaging (%)‡ | ||||
No | 265 (22) | 225 (26) | 40 (11) | |
Yes | 961 (78) | 648 (74) | 313 (89) | <0.001 |
Cyst on imaging (%)‡ | ||||
No | 1108 (90) | 780 (89) | 328 (93) | |
Yes | 118 (10) | 93 (11) | 25 (7) | 0.06 |
Median preop T1 contrast-enhancing vol in cm3(min, max)§ | 31.0 (0.3, 186.3) | 25.9 (0.3, 175.1) | 46.3 (3.9, 186.3) | <0.001 |
Values represent numbers of patients (%) unless otherwise specified. Percentages may not add up to 100 due to rounding.
Percentages in first column to the right are based on a total of 1224; second column, 873; and third column, 351.
Percentages in first column to the right are based on a total of 1226; and second column, 873.
Values are based on 1206 tumors.
Complete resection (100%) of the T1 contrast-enhancing part of the tumor was achieved in 876 patients (71%). Less than complete resection (78% to < 100%) was achieved in 353 patients (29%). These 2 patient groups had different distributions with respect to preoperative KPS scores, tumor functional grade, preoperative contrast-enhancing tumor volume, and presence of necrosis, as well as preoperative symptoms (Table 1).They were comparable in terms of patient age, sex, presence of cysts on imaging, and prior treatment status. The median duration of follow-up among the 194 patients remaining alive at the end of the study was 19.9 months (range 0.1–190.1 months).
Analysis of Overall Survival
The Kaplan-Meier estimate of median postoperative survival time for the 1229 study patients was 13.4 months (95% CI 12.6–14.1 months;Table 2).Patients with complete resection (100% of contrast-enhancing region) had a median survival time of 15.2 months (95% CI 14.1–16.3 months), a duration significantly longer than the median survival of patients with less than complete resection (9.8 months, 95% CI 8.8–10.8 months; p < 0.001;Fig. 1).Patients without complete resection had a univariate HR of 1.87 (95% CI 1.63–2.14, p < 0.001) compared with patients undergoing complete resection (Table 3).This observed difference in survival between these 2 groups remained significant when other prognostic factors were adjusted for in a multivariate Cox model analysis: less than complete resection had an adjusted HR of 1.53 (95% CI 1.33–1.77, p < 0.001). Additional factors found to have a significant negative independent effect on survival duration in the multivariate analysis included an older age at surgery (HR 1.03, 95% CI 1.02–1.03; p < 0.001), a previously treated status (HR 1.67, 95% CI 1.46–1.90; p < 0.001), a preoperative KPS score less than 70 (HR 1.30, 95% CI 1.04–1.64; p = 0.02), and a larger preoperative contrast-enhancing tumor volume (HR 1.005, 95% CI 1.002–1.007; p < 0.001). The presence of a cyst had a positive impact on the duration of survival (HR 0.63, 95% CI 0.50–0.79; p < 0.001).
Outcomes in 1229 patients with GBM treated with resection of the T1 contrast-enhancing volume*
Variable | Total No. | 100% Resection | 78% to <100% Resection | p Value |
---|---|---|---|---|
No. of patients | 1229 | 876 | 353 | |
EOR of T1 contrast-enhancing vol (%) | ||||
100% | 876 (71) | |||
78% to <100% | 353 (29) | |||
Vital status at last follow-up (%) | ||||
Alive | 194 (16) | 154 (18) | 40 (11) | |
Dead | 1035 (84) | 722 (82) | 313 (89) | 0.007 |
Median overall survival in mos (95% CI) | 13.4 (12.6, 14.1) | 15.2 (14.1, 16.3) | 9.8 (8.8, 10.8) | <0.001 |
30-day postop complications (%) | ||||
No | 950 (77) | 693 (79) | 257 (73) | |
Yes | 279 (23) | 183 (21) | 96 (27) | 0.02 |
30-day postop neurological complications (%) | ||||
No | 1002 (82) | 728 (83) | 274 (78) | |
Yes | 227 (18) | 148 (17) | 79 (22) | 0.02 |
Motor deficit | 115 (9) | 75 (9) | 40 (11) | 0.13 |
Speech impairment | 77 (6) | 50 (6) | 27 (8) | 0.20 |
Visual impairment | 33 (3) | 21 (2) | 12 (3) | 0.32 |
Seizure | 30 (2) | 17 (2) | 13 (4) | 0.07 |
Hemorrhage | 17 (1) | 12 (1) | 5 (1) | 1.00 |
Intracranial hemorrhage | 14 (1) | 10 (1) | 4 (1) | 1.00 |
Hydrocephalus | 12 (<1) | 7 (<1) | 5 (1) | 0.34 |
Cognitive/memory/mental status problems | 12 (<1) | 8 (<1) | 4 (1) | 0.06 |
Sensory deficit | 7 (<1) | 4 (<1) | 3 (<1) | 0.42 |
Headache | 8 (<1) | 5 (<1) | 3 (<1) | 0.70 |
Cranial nerve deficit | 6 (<1) | 5 (<1) | 1 (<1) | 0.68 |
Other† | 36 (3) | 24 (3) | 12 (3) | 0.54 |
Values represent numbers of patients (%) unless otherwise specified.
Other includes edema, cerebrospinal fluid leak, cerebral abscess, pneumocephalus, superior sagittal sinus thrombosis, fluid entrapment in resection cavity, left posterior cerebral artery infarct, nausea and vomiting, subdural empyema, bladder dysfunction, and gait abnormalities.
Overall survival analysis in 1229 patients with GBM
Variable | Kaplan-Meier Analysis | Univariate Analysis | Multivariate Analysis | |||||||
---|---|---|---|---|---|---|---|---|---|---|
Total No. of Patients | Events | Median Survival (mos) | 95% CI | HR | 95% CI | P Value | HR | 95% CI | P Value | |
Age at surgery | - - - - - - | - - - - - - | - - - - - - | - - - - - - | 1.03 | 1.02–1.03 | <0.001 | 1.03 | 1.02–1.03 | <0.001 |
年龄在岁手术 | ||||||||||
≤51.7 | 456 | 349 | 17.3 | 15.3–19.4 | 1.00 | |||||
>51.7 | 773 | 686 | 11.9 | 11.1–12.7 | 1.74 | 1.53–1.99 | <0.001 | |||
Sex (%) | ||||||||||
Male | 758 | 639 | 13.0 | 12.2–13.9 | 1.00 | |||||
Female | 471 | 396 | 13.7 | 12.0–15.4 | 0.88 | 0.78–1.00 | 0.04 | |||
Preop KPS score | - - - - - - | - - - - - - | - - - - - - | - - - - - - | 0.98 | 0.98–0.99 | <0.001 | |||
Preop KPS score | ||||||||||
≥70 | 1134 | 951 | 13.6 | 12.9–14.3 | 1.00 | 1.00 | ||||
<70 | 95 | 84 | 9.9 | 7.6–12.2 | 1.49 | 1.19–1.86 | <0.001 | 1.30 | 1.04–1.64 | 0.02 |
Tumor functional grade* | ||||||||||
1 | 166 | 125 | 17.8 | 15.7–19.8 | 1.00 | |||||
2 | 589 | 509 | 12.7 | 11.6–13.8 | 1.37 | 1.13–1.67 | 0.002 | |||
3 | 469 | 396 | 13.2 | 12.2–14.2 | 1.29 | 1.06–1.58 | 0.013 | |||
Necrosis on imaging† | ||||||||||
No | 265 | 195 | 17.8 | 15.4–20.3 | 1.00 | |||||
Yes | 961 | 837 | 12.7 | 11.9–13.5 | 1.60 | 1.37–1.88 | <0.001 | |||
Cyst on imaging† | ||||||||||
No | 1108 | 941 | 13.0 | 12.2–13.7 | 1.00 | 1.00 | ||||
Yes | 118 | 91 | 20.3 | 15.3–25.3 | 0.60 | 0.48–0.75 | <0.001 | 0.63 | 0.50–0.79 | <0.001 |
Symptoms before surgery | ||||||||||
No | 105 | 87 | 13.5 | 11.6–15.5 | 1.00 | |||||
Yes | 1124 | 948 | 13.4 | 12.6–14.2 | 1.05 | 0.84–1.31 | 0.65 | |||
Previously treated | ||||||||||
No | 752 | 613 | 14.9 | 13.7–16.1 | 1.00 | 1.00 | ||||
Yes | 477 | 422 | 10.8 | 9.9–11.7 | 1.42 | 1.26–1.62 | <0.001 | 1.67 | 1.46–1.90 | <0.001 |
Preop T1 contrast-enhancing vol | - - - - - - | - - - - - - | - - - - - - | - - - - - - | 1.004 | 1.002–1.006 | <0.001 | 1.005 | 1.002–1.007 | <0.001 |
EOR of T1 contrast-enhancing vol | ||||||||||
100% | 876 | 722 | 15.2 | 14.1–16.3 | 1.00 | 1.00 | ||||
78% to ≤100% | 353 | 313 | 9.8 | 8.8–10.8 | 1.87 | 1.63–2.14 | <0.001 | 1.53 | 1.33–1.77 | <0.001 |
Values based on 1224 patients,
Values based on 1226 patients.
The 2 resection groups had significant differences in important characteristics, which could have impacted their effect on overall survival.Table 4shows the effect of EOR within various categories of these characteristics (variables). The hazard ratio remained basically unaltered in terms of direction, and the p values remained significant in all categories except for an asymptomatic presentation. There were too few patients without symptoms in this series.
Survival relative to EOR within categories of prognostic variables in 1229 patients with GBM
Variables | 78% to ≤100% Resection | 100% Resection | Univariate Cox Analysis* | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Total No. of Patients | No. of Events | Median Survival (mos) | 95% CI | Total No. of Patients | No. of Events | Median Survival (mos) | 95% CI | HR | 95% CI | p Value | |
年龄在岁手术 | |||||||||||
≤51.7 | 119 | 100 | 8.9 | 6.8–11.0 | 337 | 249 | 20.6 | 17.4–23.8 | 2.47 | 1.94–3.14 | <0.001 |
>51.7 | 234 | 213 | 9.9 | 9.0–10.9 | 539 | 473 | 13.4 | 12.4–14.3 | 1.54 | 1.31–1.82 | <0.001 |
Sex | |||||||||||
Male | 219 | 194 | 9.8 | 8.5–11.2 | 539 | 445 | 14.7 | 13.6–15.7 | 1.86 | 1.57–2.21 | <0.001 |
Female | 134 | 119 | 9.7 | 8.3- - - - - -11.2 | 337 | 277 | 16.5 | 14.8–18.2 | 1.86 | 1.50–2.32 | <0.001 |
Preop KPS score | |||||||||||
≥70 | 315 | 278 | 10.0 | 8.9–11.1 | 819 | 673 | 15.5 | 14.5–16.6 | 1.85 | 1.60–2.13 | <0.001 |
<70 | 38 | 35 | 7.6 | 4.8–10.4 | 57 | 49 | 13.0 | 8.6–17.5 | 1.81 | 1.14–2.86 | 0.01 |
Tumor functional grade | |||||||||||
1 | 15 | 14 | 9.3 | 6.3–12.3 | 151 | 111 | 18.4 | 16.4–20.5 | 2.41 | 1.37–4.25 | 0.002 |
2 | 180 | 161 | 9.1 | 8.2–10.0 | 409 | 348 | 14.8 | 13.4–16.1 | 1.73 | 1.43–2.09 | <0.001 |
3 | 156 | 136 | 10.6 | 9.0–12.1 | 313 | 260 | 14.7 | 13.4–16.0 | 1.88 | 1.52–2.40 | <0.001 |
Necrosis on imaging | |||||||||||
No | 40 | 33 | 11.9 | 9.2–14.6 | 225 | 162 | 19.9 | 17.4–22.3 | 2.01 | 1.38–2.94 | <0.001 |
Yes | 313 | 280 | 9.7 | 8.8–10.6 | 648 | 557 | 14.4 | 13.5–15.4 | 1.74 | 1.50–2.01 | <0.001 |
Cyst on imaging | |||||||||||
No | 328 | 295 | 9.6 | 8.7–10.6 | 780 | 646 | 14.7 | 13.7–15.6 | 1.80 | 1.56–2.06 | <0.001 |
Yes | 25 | 18 | 11.5 | 8.6–14.4 | 93 | 73 | 23.2 | 15.7–30.6 | 2.36 | 1.37–4.07 | 0.001 |
Symptoms before surgery | |||||||||||
No | 20 | 17 | 8.8 | 0.0–17.6 | 85 | 70 | 14.6 | 11.3–17.8 | 1.49 | 0.87–2.54 | 0.15 |
Yes | 333 | 296 | 9.8 | 8.9–10.7 | 791 | 652 | 15.4 | 14.2–16.5 | 1.90 | 1.65–2.18 | <0.001 |
Previously treated | |||||||||||
No | 213 | 183 | 10.4 | 9.2–11.6 | 539 | 430 | 17.9 | 16.3–19.6 | 2.08 | 1.74–2.48 | <0.001 |
Yes | 140 | 130 | 8.3 | 7.2–9.3 | 337 | 292 | 12.4 | 11.0–13.8 | 1.60 | 1.30–1.98 | <0.001 |
Preop T1 contrast-enhancing vol in cm3 | |||||||||||
≥31.0 | 244 | 217 | 14.2 | 12.9–15.5 | 361 | 300 | 11.4 | 10.4–12.4 | 1.81 | 1.51–2.16 | <0.001 |
<31.0 | 109 | 96 | 16.1 | 14.5–17.7 | 492 | 400 | 14.6 | 13.3–15.9 | 1.72 | 1.37–2.16 | <0.001 |
100% resection is the referent group.
We performed a close and separate examination of previously untreated and previously treated patients and found similar negative prognostic effects for a less-than-complete resection in both groups (adjusted HR 1.69, 95% CI 1.41–2.04, p < 0.001, and adjusted HR 1.33, 95% CI 1.06–1.67, p = 0.02, respectively; data not shown).
To address one of the secondary objectives of the study—that is, whether additional resection of the FLAIR abnormality region beyond the enhancing area would have a significant effect on survival time overall or within defined patient subgroups—patients who underwent complete resection of the contrast-enhancing tumor volume were further examined. Of the 876 patients who had undergone complete resection, 643 underwent further resection of the T2 FLAIR abnormality region and had FLAIR data. The distribution of the EOR of the FLAIR abnormality area is presented inFig. 2. Approximately 18% of the patients had a negative EOR, reflecting the development of postoperative edema. In the 643-patient subgroup, a FLAIR abnormality resection ≥ 53.21%, compared with one < 53.21%, was significantly associated with improved survival in the univariate analysis (median 20.7 months, compared with 15.5 months; crude HR 0.68, 95% CI 0.55–0.84, p < 0.001;Tables 5and6).In the multivariate analysis, the previously treated group with a FLAIR resection < 53.21% fared significantly worse than all others (that is, previously treated patients who underwent FLAIR resection ≥ 53.21%, previously untreated patients who underwent FLAIR resection < 53.21%, and previously untreated patients who underwent FLAIR resection ≥ 53.21%). The untreated group with at least 53.21% resection had the longest survival. Prior treatment in the 643-patient subgroup included prior resection in 100% of the patients, prior radiation in 189 (79%), and prior chemotherapy in 177 (74%; 61% of which received temozolomide; 72%, both radiation and chemotherapy; and 60%, both radiation and temozolomide). The median preoperative FLAIR abnormality volume was 57.3 cm3in patients who had received prior radiation and/or chemo, compared with 33.3 cm3in those who had not (p = 0.17). The median extent of FLAIR abnormality resection was 22% in patients who received prior radiation and/or chemo compared with 32% in those who did not (p = 0.02).Figure 3显示了根据t三次采油对生存的影响reatment group. A younger age, a higher preoperative KPS score, a cystic tumor, and a smaller preoperative contrast-enhancing tumor on imaging were strong independent positive prognostic factors in this analysis. All categorized continuous variables had a p value < 0.001. Those p values remained significant after adjustment for an approximate true false-positive rate of 10% based on the method of Altman et al.1
Survival in 643 patients with GBM whose entire T1 contrast-enhancing tumor portion was resected and who had EOR of T2 FLAIR abnormality data*
Variable | No. of Patients w/ Data Available | No. of Death Events | Median Survival (mos) | 95% CI | p Value (log-rank) | Univariate Analysis | Multivariate Analysis | ||||
---|---|---|---|---|---|---|---|---|---|---|---|
HR | 95% CI | P Value | HR | 95% CI | P Value | ||||||
Overall survival | 643 (100) | 502 | 16.8 | 15.4–18.2 | |||||||
Age at surgery | - - - - - - | - - - - - - | - - - - - - | - - - - - - | - - - - - - | 1.03 | 1.02–1.04 | <0.001 | 1.03 | 1.02–1.04 | <0.001 |
Sex | |||||||||||
Male | 399 (62) | 315 | 15.2 | 13.8–16.7 | 1.00 | ||||||
Female | 244 (38) | 187 | 18.7 | 16.8–20.5 | 0.003 | 0.76 | 0.63–0.91 | 0.004 | |||
Preop KPS score | - - - - - - | - - - - - - | - - - - - - | - - - - - - | - - - - - - | 0.98 | 0.98–0.99 | <0.001 | 0.988 | 0.981–0.995 | 0.001 |
Preop KPS score | |||||||||||
≥70 | 601 (93) | 468 | 16.9 | 15.5–18.3 | 1.00 | ||||||
<70 | 42 (6) | 34 | 14.7 | 9.6–19.9 | 0.14 | 1.30 | 0.92–1.84 | 0.14 | |||
Tumor functional grade | |||||||||||
1 | 127 (20) | 88 | 18.9 | 17.2–20.5 | 1.00 | ||||||
2 | 304 (47) | 249 | 15.5 | 13.9 - -17.2 | 1.29 | 1.01–1.65 | 0.04 | ||||
3 | 212 (33) | 165 | 16.9 | 13.7–20.1 | 0.03 | 1.03 | 0.80–1.34 | 0.81 | |||
Necrosis on imaging | |||||||||||
No | 184 (29) | 129 | 19.9 | 18.0–21.8 | 1.00 | ||||||
Yes | 459 (71) | 373 | 15.4 | 14.0–16.7 | 0.003 | 1.35 | 1.10–1.65 | 0.004 | |||
Cyst on imaging | |||||||||||
No | 575 (89) | 449 | 15.9 | 14.6–17.2 | 1.00 | ||||||
Yes | 68 (10) | 53 | 24.5 | 14.5–34.5 | 0.007 | 0.68 | 0.51–0.90 | 0.008 | 0.66 | 0.49–0.90 | 0.009 |
Symptoms before surgery | |||||||||||
No | 64 (10) | 50 | 16.9 | 13.9–19.9 | 1.00 | ||||||
Yes | 579 (90) | 452 | 16.7 | 15.2–18.1 | 0.96 | 0.99 | 0.74–1.33 | 0.96 | |||
Previously treated | |||||||||||
No | 403 (63) | 302 | 19.7 | 18.0–21.4 | 1.00 | ||||||
Yes | 240 (37) | 200 | 12.4 | 10.7–14.0 | <0.001 | 1.69 | 1.41–2.02 | <0.001 | |||
Preop T1 contrast-enhancing vol | - - - - - - | - - - - - - | - - - - - - | - - - - - - | - - - - - - | 1.004 | 1.001–1.007 | 0.02 | 1.004 | 1.001–1.008 | 0.03 |
Preop T2 FLAIR abnormality vol | - - - - - - | - - - - - - | - - - - - - | - - - - - - | - - - - - - | 1.000 | 0.999–1.002 | 0.69 | |||
EOR of T2 FLAIR abnormality vol | |||||||||||
<53.21% | 484 (75) | 390 | 15.5 | 14.1–17.0 | 1.00 | ||||||
≥53.21% | 159 (25) | 112 | 20.7 | 18.2–23.1 | <0.001 | 0.68 | 0.55–0.84 | <0.001 | |||
EOR of T2 FLAIR abnormality vol by previous treatment | |||||||||||
Treated, <53.21% | 199 (31) | 166 | 11.2 | 9.8–12.6 | 1.00 | 1.00 | |||||
Treated, ≥53.21% or not treated, <53.21% | 327 (51) | 258 | 18.7 | 17.0–20.4 | 0.62 | 0.51–0.75 | <0.001 | 0.49 | 0.40–0.61 | <0.001 | |
Not treated, ≥53.21% | 117 (18) | 78 | 23.0 | 17.6–28.4 | <0.001 | 0.45 | 0.34–0.59 | <0.001 | 0.43 | 0.33–0.57 | <0.001 |
Values represent numbers of patients (%) unless otherwise specified.
Survival relative to EOR of T2 FLAIR abnormality in 643 patients with GBM who had the entire contrast-enhancing portion of their tumor resected
EOR of T2 FLAIR Abnormality* | Total No. of Patients | Total No. of Events | Median Survival (mos) | 95% CI |
---|---|---|---|---|
< −50% | 35 | 32 | 15.9 | 11.2–20.1 |
−50% to <−10% | 44 | 37 | 16.2 | 10.1–22.3 |
−10% to <0 | 40 | 37 | 10.5 | 8.0–12.9 |
0 to <10% | 62 | 47 | 16.1 | 12.2–20.1 |
10% to <20% | 78 | 61 | 14.6 | 9.3–19.9 |
20% to <30% | 72 | 65 | 14.6 | 12.2–16.9 |
30% to <40% | 69 | 55 | 16.0 | 13.1–18.9 |
40% to <50% | 62 | 41 | 16.2 | 10.7–21.8 |
50% to <60% | 43 | 26 | 26.4 | 9.9–42.8 |
60% to <70% | 40 | 32 | 19.5 | 11.4–27.7 |
70% to <80% | 26 | 19 | 22.5 | 11.4–33.5 |
80% to <90% | 33 | 23 | 20.2 | 16.5–23.9 |
90% to 100% | 39 | 27 | 20.2 | 15.7–24.7 |
Negative values indicate that the T2 FLAIR abnormality actually increased after surgery relative to the preoperative value; positive values, that it decreased.
Postoperative Complications
A review of postoperative outcomes revealed that 23% of the 1229 patients included in this study suffered from postoperative complications and 18% from neurological complications (Table 2).最常见的神经系统并发症motor deficits (9%), followed by speech impairments (6%) and visual deficits (3%). Seizures were observed in 2% of the patients.
The group with less than complete resection had a significantly higher rate of complications than the complete resection group (27% vs 21%, respectively, p = 0.02) and a significantly higher rate of neurological complications (22% vs 17%, p = 0.02).
In the complete resection group, we compared the overall and neurological postoperative complication rates between patients who also underwent ≥ 53.21% FLAIR resection with those who had < 53.21% resection. The rates of overall complications were significantly higher in patients with FLAIR resection < 53.21% (26% vs 18%, p = 0.04;Table 7).The rates of neurological complications were not significantly different between the 2 groups (p = 0.12).
Outcomes in 643 patients with GBM treated with additional resection of T2 FLAIR abnormality region*
Variable | Total No. | ≥53.21% Resection | <53.21% Resection | p Value |
---|---|---|---|---|
No. of patients | 643 | 159 | 484 | |
30-day postop complications (%) | ||||
No | 490 (76) | 131 (82) | 359 (74) | |
Yes | 153 (24) | 28 (18) | 125 (26) | 0.04 |
30-day postop neurological complications (%) | ||||
No | 523 (81) | 136 (86) | 387 (80) | |
Yes | 120 (19) | 23 (14) | 97 (20) | 0.12 |
Motor deficit | 57 (9) | 8 (5) | 49 (10) | 0.05 |
Speech impairment | 46 (7) | 9 (6) | 37 (8) | 0.40 |
Visual impairment | 18 (3) | 4 (3) | 14 (3) | 1.00 |
Seizure | 12 (2) | 1 (<1) | 11 (2) | 0.31 |
Hemorrhage | 11 (2) | 1 (<1) | 10 (2) | 0.31 |
Intracranial hemorrhage | 9 (1) | 1 (<1) | 8 (2) | 0.46 |
Hydrocephalus | 7 (1) | 1 (<1) | 6 (1) | 1.00 |
Cognitive/memory/mental status problems | 8 (1) | 3 (2) | 5 (1) | 0.42 |
Sensory deficit | 4 (<1) | 1 (<1) | 3 (<1) | 1.00 |
Headache | 3 (<1) | 1 (<1) | 2 (<1) | 0.57 |
Cranial nerve deficit | 4 (<1) | 1 (<1) | 3 (<1) | 1.00 |
Other† | 22 (3) | 3 (2) | 19 (4) | 0.22 |
The value 53.21% comes from CART (classification and regression tree) analysis, also called recursive partitioning.
Other includes edema, cerebrospinal fluid leak, pneumocephalus, superior sagittal sinus thrombosis, fluid entrapment in resection cavity, posterior cerebral artery infarct, nausea and vomiting, subdural empyema, bladder dysfunction, and gait abnormalities.
Discussion
The current standard of care for patients with GBM is resection followed by adjuvant chemoradiotherapy. Several patient and tumor variables have been shown to affect the prognosis of patients with GBM, including age, preoperative functional status, and preoperative imaging characteristics of the tumor. Additionally, the prognostic value of EOR has been repeatedly confirmed,15,19,20and authors of these studies and others have suggested that at least 78%–89% of the contrast-enhancing tumor volume needs to be resected for a significant survival advantage. In the current study, we first explored the influence of complete resection, that is, surgical removal of the entire contrast-enhancing region, on patient survival. Complete resection was achieved in 876 patients (71%). The 15.2-month median survival time among these patients was about 5.4 months longer than the median survival for patients undergoing less than complete resection (p < 0.001), a finding supported by our multivariate analysis (adjusted HR for the less than complete resection group 1.53, 95% CI 1.33–1.77, p < 0.001). The 2 resection groups had significant differences in important characteristics, which could have affected overall survival. We examined the effect of EOR within various categories of these characteristics. The hazard ratio remained basically unaltered in terms of direction, and the p values remained significant in all categories except for an asymptomatic presentation, alleviating concern that such a finding is attributable to a different distribution of prognostic variables between the 2 resection groups.
Therefore, the conclusion supported by the data and the one that should be emphasized to practicing neurosurgeons is that every attempt should be made to resect the totality of the contrast-enhancing portion of a GBM tumor, as this approach will lead to a maximal survival benefit. In addition, given that the contrast-enhancing portion of a tumor is entirely composed of tumor cells,16the goal of complete resection should be achievable in most patients with GBM, even when the lesion is located within or near eloquent brain.
并发症的发生率并不高patients who underwent complete resection than in those who underwent less than complete resection. Rather, it was the other way around. This finding probably reflects the sicker status of the latter group, which we corrected for in the multivariate analysis of survival.
As noted previously, the invasiveness of GBM has been demonstrated in vitro, in vivo, and in postmortem and other studies. The majority of patients succumbing to GBM die of local tumor recurrence, presumably from re-growth of infiltrating tumors left behind in the vicinity of the original tumor. Cerebral edema, demyelination, and/or surgery-related injury represent an important component of the T2 FLAIR abnormality region.9,28In cases in which edema is a significant component after resection of the enhancing mass, the FLAIR signal shrinks considerably. But the tumor component reflecting the invasive nature of GBM cannot be ignored.8,9,13,26,28The area outside the contrast-enhancing region on T1-weighted MRI is usually infiltrated by tumor cells. These appear on T2 FLAIR images;8,13therefore, we explored the effect of additional resection of the T2 FLAIR abnormality region in the subgroup that had undergone complete resection of their contrast-enhancing tumor.
We identified the subset of patients in whom some FLAIR abnormality region was resected. Of note, only around 30% of patients in our study had > 50% resection of the T2 FLAIR abnormality. When these cases were performed, there was no evidence in the literature to support a resected FLAIR region as a significant positive prognostic factor, nor was there evidence on the safety of additional FLAIR resection, hence the apparent reluctance of our surgeons to push the boundary of T2 FLAIR resection. To our knowledge, this study is the first to present data on the effect of T2 FLAIR resection in a large number of patients. The literature does support aggressive resection of the enhancing lesion, a practice rigorously followed at our center, as reflected in our study data on resecting the contrast-enhancing tumor volume. Resecting a portion of the T2 FLAIR abnormality region did confer an added survival benefit over complete resection at the univariate analysis level. The trend remained strong at the multivariate level. Such a benefit was observed in patients undergoing at least 53.21% resection of the FLAIR abnormality compared with those undergoing a less extensive resection. This advantage of the ≥ 53.21% EOR group over the < 53.21% EOR group was initially observed in the univariate analysis in both the previously untreated patient subset (median survival 23.2 months, 95% CI 17.8–28.6 months compared with 18.7 months, 95% CI 16.9–20.5 months, respectively) and in the previously treated patient subset (median survival 19.8 months, 95% CI 15.3–24.3 months compared with 11.2 months, 95% CI 9.8–12.6 months, respectively). In the multivariate analysis, after correcting for other independent prognostic factors, the previously treated group with < 53.21% resection had significantly shorter survival than the other 3 categories (that is, previously treated patients who underwent FLAIR resection ≥ 53.21%, previously untreated patients who underwent FLAIR resection < 53.21%, and previously untreated patients who underwent FLAIR resection ≥ 53.21%). The previously untreated group with ≥ 53.21% resection had the longest survival. (Of note, patients previously treated with radiation and/or chemotherapy had a significantly less extensive T2 FLAIR abnormality resection than those who had not been previously treated.) Therefore, resection of a significant portion of the T2 FLAIR abnormality region, if feasible and safely attempted, could have a beneficial impact on the survival of patients with GBM. We believe these results have significant value. Future reviews at our and other centers will hopefully shed more light on the issue, with confirmatory evidence and a better definition of the thresholds involved. Calls to maximize the EOR raise concerns about associated increases in patient morbidity.22We reviewed the surgical morbidity associated with resection in our series. Complete resection was found to be relatively safe if it was feasible, as was removal of the T2 FLAIR abnormality region outside of the T1 contrast-enhancing GBM region. The decreased risk probably reflects the increased use of intraoperative neurophysiological monitoring; functional cortical, sub-cortical, and language mapping; awake craniotomy technique; neuronavigation overlapping with diffusion tensor imaging (DTI) and functional imaging; and intraoperative MRI. The focus of any GBM resection should be maximal tumor removal while protecting against postoperative neurological or other deficits. If complete (100%) resection of the contrast-enhancing tumor is not achievable, then less extensive resections (as dictated primarily by functional mapping) may still provide both symptomatic and survival benefits. Neurosurgeons should aim to perform the maximum resection possible without hurting the patient. Additional resection of the T2 FLAIR region appears to confer an added survival advantage without added risk.
本研究有很多局限性,主啊ne being that the compared cohorts may be inherently different. There were significant differences in baseline characteristics between patients undergoing 100% and less than 100% resection of the contrast-enhancing lesion. A closer examination showed that the findings remained basically unaltered when the effects of EOR on survival were compared within subgroups of these baseline characteristics. In addition to the known differences between the groups being compared, there are probably other inherent differences that are not necessarily easily controlled for within such a study design. In addition, although most of the data were collected prospectively, the study is a retrospective one. As a result, important information was not available to review—for example, the specific reason the neurosurgeon chose to resect the volume that he or she resected—as well as the unavailability of T2 FLAIR MRI studies for some of the earlier patients, which limited the power of the study in some of the subgroup analyses and possibly introduced bias. Despite these limitations, most of the data pointed in the same direction, lending support to the legitimacy of the conclusions derived from this study.
In the absence of better imaging techniques to distinguish between tumor and cerebral edema on FLAIR images, we retrospectively assessed the relation between resection of the FLAIR abnormality area and survival duration and concluded that a greater EOR of the FLAIR area is probably associated with longer patient survival.
Conclusions
根据是什么,据我们所知,最大的硅ngle-center series of GBM patients with extensive tumor resections, this study supports the established association between EOR and survival. It also provides evidence that the maximum safe resection of contrast-enhancing GBM volume results in a significant survival advantage relative to less extensive resections. Such an outcome is achieved without a significant increase in postoperative neurological deficits or other complications. In addition, in the subgroup that underwent complete resection of the T1 contrast-enhancing region, additional resection of the T2 FLAIR abnormality region may confer a significant added survival advantage. We conclude that pushing the boundary to 100% resection and beyond, along with the removal of a significant amount of the FLAIR abnormality region, may result in longer survival without significant increases in postoperative morbidity and should be further investigated.
Acknowledgments
We thank David M. Wildrick, PhD, for editorial assistance and Stephanie Jenkins for assistance with manuscript preparation. Data for this work were compiled from a database partly funded by an institutional development grant from The University of Texas MD Anderson Cancer Center.
Author Contributions
Conception and design: Sawaya, Li, Suki. Acquisition of data: Sawaya Li, Suki. Analysis and interpretation of data: all authors. Drafting the article: Li, Suki. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Sawaya. Statistical analysis: Suki. Administrative/technical/material support: Sawaya. Study supervision: Sawaya, Suki.
References
-
1 ↑
AltmanDG,,LausenB,,SauerbreiW,&SchumacherM:Dangers of using “optimal” cutpoints in the evaluation of prognostic factors.J Natl Cancer Inst86:829- - - - - -835,1994
-
2 ↑
BernsteinJJ,,GoldbergWJ,,LawsERJr,,CongerD,,MorrealeV,&WoodLR:C6 glioma cell invasion and migration of rat brain after neural homografting: ultrastructure.开云体育app官方网站下载入口26:622- - - - - -628,1990
-
3 ↑
BurgerPC,,DuboisPJ,,ScholdSCJr,,SmithKRJr,,OdomGL,&CraftsDC,et al.:Computerized tomographic and pathologic studies of the untreated, quiescent, and recurrent glioblastoma multiforme.J Neurosurg58:159- - - - - -169,1983
-
4 ↑
BurgerPC,,HeinzER,,ShibataT,&KleihuesP:Topographic anatomy and CT correlations in the untreated glioblastoma multiforme.J Neurosurg68:698- - - - - -704,1988
-
5 ↑
CerameMA,,GuthikondaM,&KohliCM:Extraneural metastases in gliosarcoma: a case report and review of the literature.开云体育app官方网站下载入口17:413- - - - - -418,1985
-
6 ↑
ChoucairAK,,LevinVA,,GutinPH,,DavisRL,,SilverP,&EdwardsMS,et al.:Development of multiple lesions during radiation therapy and chemotherapy in patients with gliomas.J Neurosurg65:654- - - - - -658,1986
-
7 ↑
DolecekTA,,ProppJM,,StroupNE,&KruchkoC:CBTRUS statistical report: primary brain and central nervous system tumors diagnosed in the United States in 2005–2009.Neuro Oncol14:Suppl 5v1- - - - - -v49,2012
-
8 ↑
EarnestFIV,,KellyPJ,,ScheithauerBW,,KallBA,,CascinoTL,&EhmanRL,et al.:Cerebral astrocytomas: histopathologic correlation of MR and CT contrast enhancement with stereotactic biopsy.Radiology166:823- - - - - -827,1988
-
9 ↑
EbosJM,,LeeCR,&KerbelRS:Tumor and host-mediated pathways of resistance and disease progression in response to antiangiogenic therapy.Clin Cancer Res15:5020- - - - - -5025,2009
-
10 ↑
GreeneGM,,HitchonPW,,SchelperRL,,YuhW,&DysteGN:Diagnostic yield in CT-guided stereotactic biopsy of gliomas.J Neurosurg71:494- - - - - -497,1989
-
11 ↑
HochbergFH,&PruittA:Assumptions in the radiotherapy of glioblastoma.Neurology30:907- - - - - -911,1980
-
12 ↑
IbayashiN,,HermanMM,,BoydJC,&RubinsteinLJ:Relationship of invasiveness to proliferating activity and to cyto-skeletal protein production in human neuroepithelial tumors maintained in an organ culture system: use of human cortex and dura as supporting matrices.开云体育app官方网站下载入口26:629- - - - - -637,1990
-
13 ↑
KellyPJ,,Daumas-DuportC,,KispertDB,,KallBA,,ScheithauerBW,&IlligJJ:Imaging-based stereotaxic serial biopsies in untreated intracranial glial neoplasms.J Neurosurg66:865- - - - - -874,1987
-
15 ↑
LacroixM,,Abi-SaidD,,FourneyDR,,GokaslanZL,,ShiW,&DeMonteF,et al.:A multivariate analysis of 416 patients with glioblastoma multiforme: prognosis, extent of resection, and survival.J Neurosurg95:190- - - - - -198,2001
-
16 ↑
LangFF,,SawayaR,,SukiD,,McCutcheonIE,&HessKR:Glioblastoma resection.J Neurosurg116:1166- - - - - -1167,author reply 1167–1168, 2012
-
17 ↑
LeiferD,,MooreT,,UkenaT,,WilnerD,,ThorA,&Hedley-WhyteET:与肝转移灶状胶质母细胞瘤the absence of surgery. Case report.J Neurosurg71:772- - - - - -776,1989
-
18 ↑
LiwniczBH,&RubinsteinLJ:The pathways of extraneural spread in metastasizing gliomas: a report of three cases and critical review of the literature.Hum Pathol10:453- - - - - -467,1979
-
19 ↑
MarkoNF,,WeilRJ,,SchroederJL,,LangFF,,SukiD,&SawayaRE:Extent of resection of glioblastoma revisited: personalized survival modeling facilitates more accurate survival prediction and supports a maximum-safe-resection approach to surgery.J Clin Oncol32:774- - - - - -782,2014
-
20 ↑
SanaiN,,PolleyMY,,McDermottMW,,ParsaAT,&BergerMS:An extent of resection threshold for newly diagnosed glioblastomas.J Neurosurg115:3- - - - - -8,2011
-
21 ↑
SawayaR,,HammoudM,,SchoppaD,,HessKR,,WuSZ,&ShiWM,et al.:Neurosurgical outcomes in a modern series of 400 craniotomies for treatment of parenchymal tumors.开云体育app官方网站下载入口42:1044- - - - - -1056,1998
-
22 ↑
StummerW,,TonnJC,,MehdornHM,,NestlerU,,FranzK,&GoetzC,et al.:Counterbalancing risks and gains from extended resections in malignant glioma surgery: a supplemental analysis from the randomized 5-aminolevulinic acid glioma resection study. Clinical article.J Neurosurg114:613- - - - - -623,2011
-
23 ↑
Van MeirEG,,HadjipanayisCG,,NordenAD,,ShuHK,,WenPY,&OlsonJJ:Exciting new advances in neuro-oncology: the avenue to a cure for malignant glioma.CA Cancer J Clin60:166- - - - - -193,2010
-
24 ↑
VertosickFTJr,&SelkerRG:幕上的脑干和脊髓转移glioblastoma multiforme: a clinical series.开云体育app官方网站下载入口27:516- - - - - -522,1990
-
25 ↑
WallnerKE,,GalicichJH,,KrolG,,ArbitE,&MalkinMG:Patterns of failure following treatment for glioblastoma multiforme and anaplastic astrocytoma.Int J Radiat Oncol Biol Phys16:1405- - - - - -1409,1989
-
26 ↑
WilsonCB:Glioblastoma: the past, the present, and the future.Clin Neurosurg38:32- - - - - -48,1992
-
27 ↑
YungWA,,HortenBC,&ShapiroWR:Meningeal gliomatosis: a review of 12 cases.Ann Neurol8:605- - - - - -608,1980
-
28 ↑
ZinnPO,,MahajanB,,SathyanP,,SinghSK,,MajumderS,&JoleszFA,et al.:Radiogenomic mapping of edema/cellular invasion MRI-phenotypes in glioblastoma multiforme.PLoS One6:e25451,2011.Erratum inPLoS One7:10.1371/annotation/b5267cb3-6aa7-47fc-a648-47f30a7cff3e, 2013