ANALYSIS OF VARIABLES AFFECTING SURVIVAL OF PATIENTS WITH ASTRACYTOMAS

208Chinese Journal of Cancer Research 16(3):208-211, 2004ANALYSIS OF VARIABLES AFFECTING SURVIVAL OF PATIENTSWITH ASTRACYTOMASWU Zhi-min吴志敏,WU Tao吴涛，YUAN Xian-hou袁先厚,CHEN Wei-guo陈卫国,JAING Pu-cha江普查Department of Neurosurgery, Zhongnan Hospital, Wuhan Universiry, Wuhan 430071CLC number: R739.41Document code: A Article ID: 1000-9604(2004)03-0208-04ABSTRACTtumor. Because these variables are interrelated, ananalysis is required to determine the variables thatObjective: To evaluate the factors that affect the influence the survival of those patients withsurvival of patients with astrocytomas. Methods: We astrocytomas.reviewed the clinical and radiological features of 104patients operated during 1995~ 2000. The features wereevaluated with univariate and multivariate analysis.MATERIALS AND METHODSResults: Univariate statistical analysis revealed thattumor localization, enhancement, edema, tumor Patientsinvasion, seizure as the first symptom recurrence, ede ma,tumor invasion， seizure as the first symptom andWe performed a retrospective study of 123recurrence of astrocytomas significantly affected the patients who were surgically treated for astrocytomassurvival, In multivariate analysis four factors showed at our institution between January 1995 and Decembersignificant danger to long survival: necrosis, 2000. After exclusion of cases with incompletepathological grade, Karnofsky Performance Scale (KPS) records， those lost in follow-up or those whosescore before operation, extent of resection. The rest pathologic material was not retrieved for review, 104factors appeared to be of no benefit to survival. patients were retained, including 56 males and 48Conclusion: The important factors that affect the long females. Tumors were classified as Grade I to Gradesurvival of patients with astrocytomas are necrosis,IV astrocytomas by the WHO classification.pathological grade, KPS score before operation and Pathological review confirmed the histologicalextent of resection.diagnosis in all patients.The tumor registry enabled us to ascertain theKey word: Astrocytoma; Survival; Prognosis; Follow-upstatus of all patients through sources such as letters orphone calls to patients and their families.Astrocytomas, the most common brain tumors, are Clinical Dataof peculiar interest because they are credited with arelative bad prognosis and short survival. SeveralWe reviewed the medical records to obtainvariables affect the survival of patients with information of the patients and clinical data. Theastrocytomas, includingage,preoperative patient's age， sex， incidence of seizures as firstperformance status according to KPS, tumor location symptoms, and KPS score at the time of presentationand preoperative CT or MRI characteristics of the were noted, we also recorded whether the patients hadoriginal or recurrent tumor at the time of presentationReceived date: Mar. 28, 2004; Accepted date: Jun. 30, 2004at our institution. All patients underwent a surgical*Author to whom correspondence should be addressed.procedure with the aim of removing as much tumorPhone: (0086-27)-67813118tissue as was consistent with the neurological functionBiography: WU Zhi-min (1976 ), male, candidate for doctor of preserved.中国煤化工section wasmedicine, Zhongnan Hospital, Wuhan University, majors in theclassified int5l resection,basic and clinical study of gliomas.subtotal resecTMYHCNMHGandbiopsyE-mail: minminye101 @ yahoo.com.cnsampling. The extent of resection is divided into four.Chinese Journal of Cancer Research 16(3):208-211, 2004_209categories based on the surgeon's operative estimates Kaplan-Meier method from date of diagnosis to dateand the postoperative CT scans or MR images. All the of death or to last known date alive. The log-rank testpatients underwent radiation therapy in addition to was usesurgical resection.difference noted in survival time. To analyze survivalrate, we used the multivariate Cox model. PotentialRadiological Studyprognostic factors were studied first in a univariatesecondly in a multivariate analysis, using a .The preoperative CT scans or MR images were stepwise selection of covariates. Significance was setreviewed prospectively in all patients and several at the level of P<0.01.tumor image characteristics were identified.The anatomical localization of the tumor wabased on radiological images and/or operativeRESULTSfindings. Tumors were classified as being locateddeeply, which involved the thalamus or basal ganglia, Univariate Analysisor superficially which involved only the cortex.Tumor necrosis was observed on CT scans or MRSingle variable corresponding to factors possiblyimages using the methods of Dean MD, et al.. The involved in the prognosis was analyzed. Amongdegree of mass effect, surrounding edema, degree of factors depending on the patient or tumor, only tencontrast enhancement， heterogeneity of contrast were influential: Grade, seizures as first symptom,enhancement, and the lobes involved were also preoperative KPS score， extent of resection,measured. .recurrence, tumor location, necrosis, edema, degree ofcontrast enhancement, and lobes involved.Statistical AnalysisThe other analyzed variables, namely, sex, age,mass effect, and heterogeneity of enhancement, wereFollowing-up information was obtained for all found to have no influence on survival time (Table 1patients. Survival curves were estimated with theand 2).Table I. Image characteristics of 104 patients with astrocytomasCharacteristicPatientsTumor locationSuperficial56DeepNecrosisNone apparentModerate41ObviousSevere5Mass effect21Minimal midline shift (< 0.5 cm)46Moderate midline shift (0.5-1 cm)Significant midline shift(> 1 cm)Edema10Less than tumor volumeApproximately equal to tumor volumeGreater than tumor volume13Degree of enhancement3Heterogeneity of enbancementEnhancement on the borderEnhancement petalled29Lobes involvedOnly one lobe中国煤化工ssTwo lodesMYHCNMHG36Three lobes.210Chinese Journal of Cancer Research 16(3):208-211, 2004Table 2. Clinical characteristics of patients with astrocytomasOrganization, is the pr inciple variable of survival timeof patients with astrocytomas, confirming the findingsCharacteristicPatientsPatients with a high- gradeSexMale56astrocytomas had a shorter survival time. The survivalFemale18time of the patients with low-grade (I or II)Age (y)<4017astrocytomas was longer than that of the patients with40-5942high- gade (III or IV) astrocytomas, and the difference≥6015was significant (P=0.001).GradeGrade I20Grade II85Table 3. Significant mutivariate predictors of suvival in30patients with astrocytomasGrade IV19Seizures as first symptomsYes27Variable_P valueRelative riskNo0.00014.527Preoperative KPS score<70140.00801.005≥7050Extent of tumor resection0.00033.952 .Extent of resectionGross-total39Sub total33Necrosis0.00601.003Partial resection23Biopsy samplingFew other series mentioned the usefulness ofRecurrence21seizures as first symptoms in patients with83be univariate predictor of longer survival time. But itis not found to be independent predictor of survival inMultivariate Analysisthe multivariate analysis. Our results suggest thatDespite the small number of patients, variablesseizures as first symptoms are important to patientsdepending on the patient or tumor were introducedwith astrocytomas. It may promote earlier diagnosisinto the Cox model. The patient's tumor location,of astrocytomas, and an earlier treatment may prolongseizures as first symptom, recurrence, edema, degreethe survival time of patients.As in other series(6-81, preoperative performanceof contrast enhancement, and lobes involved did notstatus according to the KPS is an independentappear to have a significant effect on survival. Onlypredictor of survival time. Similarly， we foundGrade, preoperative KPS score, necrosis, and extentsignificant difference in survival time betweenof tumor resection were found to be independentpatients with preoperative KPS score of more than 70predictors of survival in multivariate analysisand those with less than 70 (P=0.002).(Figurel, Table 3).The usefulness of surgery in the management ofastrocytomas is beyond disputed. Almost all authors1.2find advantages in removing them and emphasize arelationship between the survival time and the extent0.8of resection!9, 10. According to cytokinetics, evenonly 1% astrocytoma tissue was left after operation,0.6the astrocytoma tissue would recover in eightweeks-I. It has been reported that the patients gc.2 tlong survival time by aggressive resection of tumorl121Thus, surgery, although unable to cure the patients,4030 100 (m)may，because of the slow-growing tumors， maySurvival timeprovide a relative long survival time, provided thatresection is large enough and the patient'sFig. 1. Curve based on the average of covariatesneurological status is preserved enough.The recurrences of surgically treated astrocytomawere associated with a significant shorter survivalDISCUSSIONtime at the unil中国煤化工multivariatelevels. It hasival time ofOur results indicate that tumor Grade， the patients, whoI YHCNMH Gas, was notclassification developed by the World Health associated with Grade' 131. Some biological changes.Chinese Journal of Cancer Research 16(3):208-211, 2004211take place in recurrent astrocytomas, even the GradeNeurosurgical management of low-grade astrocytoma ofis the same compared with original tumor. It makes .the cerebral hemispheres []. J Neurosurg1984;recurrent tumor more invasiveness and mor61:665-73.infiltrative, causing a shorter survival time of the [4] Wakimoto H, Aoyagi M, Nakayamka T, et al. Prognosticpatients.significance of Ki-67 labeling indices obtained usingn addition to the extent of tumor resection,MIB-1 monoclonal antibody in patients withsupratentorial astrocytomas [J].Cancerscam and MR imaging characteristics were also77:373-80.associated with survival time in our series. In the [5] Burger PC, Vogel FS, Green SB, et al. Glioblastomacurrent series, patients with a deep-seated tumor had amuliforme and anaplastic astrocytoma: pathologiccriteria and prognostic implications [J]. Cancer 1985;not surprising， because surgery is considered a56:1106-11.valuable treatment of astrocytomas. The tumors lying [6] Walker MD, Alexander E Jr, Hunt WE, et al. Evaluationin the basal ganglia or hypothalamus may be so)f BCNU and /or radiotherapy in the treatment ofdangerous to remove that their resection banaplastic gliomas. A cooperative clinical trial [J]. Jconventional open surgery is often limited to a simpleNeurosurg 1978; 49:333-43.biopsy.Yeh SA， Leung SW, Sun LM，et al. PostoperativeThe absence of necrosis on imaging studies wasradiotherapy for supratentorial malignant gliomas [J]. Jan important prognostically favorable variable in thisNeurooncol 1999, 42:183-7.series, confirming the finding of Hammond, et a1.4. [8] Halperin EC, Gaspar L, Imperato J, et al. An analysis ofThe necrosis areas with in an astrocytoma are aradiotherapy data from the CNS cancer consortium' scommon imaging feature and one believed to indicaterandomized prospective trial comparing AZQ to BCNUrapid growth malignant behavior, and degeneration.in the treatment of patients with primary malignant brainAn increasing degree of enhancement of thetumors. The CNS cancer consortium []. Am J Clintumor nodule was also an independent predictor ofOncol 1993; 16:277-83.shorter survival duration at univariate level. This can [9] Philippon JH, Clemenceau SH, Fauchon FH, et al.be explained by the fact that the blood-bran barrierSupratentorial low grade astrocytomas in adults [J].must be di srupted for contrast agents to accumulateNeurosurgery 1993; 32:554 -59.within the tumor mass. The histopathological [10] Janny P, Cure H, Mohr M, et al. Low grade supratentorialcorrelation of tumor enhancement is pathologi calastrocytomas. Management and prognostic factors [J].neovascularization,endothelialproliferation,Cancer 1994; 73:1937-45.heteromorphism，infiltration, degeneration and [11] Hall PA, Levison DA. Review: assessment of cellmultiformityI5, 161.proliferation in histological material []. J Clin Patholn summary, the important factors that affect the1990; 43:184-92.ong survival of patients with astrocytomas are:12] Berger MS, Edwards MS, Wara WM, et al. Primarynecrosis，pathological grade， KPS score beforecerebral neuroblastoma. Long-term follow-up review andoperation and extent of resection. And tumortherapeutic guidelines [].Neurosurg 1983;localization，enhancement, edema， tumor inv asion,59:418-23.seizure as the first symptom, recurrence of13] NakamuraM，Konishi N， Tsunoda S， et al.astrocytoma can affect the survival at some degree.Retinoblastoma protein expression and MIB-1 correlatewith survival of patients with malignant astrocytoma [J].Cancer 1997; 80:242-9.REFERENCES[14] Hammoud MA, Sawaya R, Shi W, et al. Prognosticsignificance of preoperative MRI scans in glioblastoma1] McCormack BM, Miller DC, Budziovich GN, et al.multiforme [J]. J Neurooncol 1996; 27:65-73.Treatment and survival of low-grade astrocytoma in [15] Steen RG. Edema and tumor perfusion: characterizationadults--1977-1988 [J]. Neurosurgery 1992; 31by quantitative 1H MR imaging [D]. Am J Roentgerol636-42.1992; 158:259-64.[2Soffetti R, Chio A, Giordana MT, et al. Prognostic [16] Earnest F 4", Kelly PJ, Scheithauer BW, et al. Cerebralfactors in well-differentiated cerebral astrocytomas inastrocytomas: Histopathologic correlation of MR and CTadult [J]. Neurosurgery 1989; 24:686-92.contrast enhancement with stereotactic biopsy [J].[3] Laws ER J.， Taylor WF， Clifton MB， et al.Radiology中国煤化工YHCNMH G.