Surgical results of 158 petroclival meningiomas with special focus on standard craniotomies

The appropriate treatment of PCM is a matter of debate [2,3,4,5,6,7,8,9,9,10,11,12,13,14]. The risk of postoperative neurological deficits is high [2, 13, 19,20,21,22,23,24,25,26,27,28]. Besides surgery alternative treatments are considered like radiation or observation [29,30,31,32]. Since many of these lesions cause brainstem compression, surgical decompression is indicated, although the symptoms of the patients might only be mild. Over the last decades, the indication for surgery changed from aggressive GTR to subtotal or even “wait and scan” policies [13, 33]. Data about the natural course of the tumors are sparse. In 2003, van Havenbergh et al. concluded that the growth pattern of these tumors is unpredictable and that in relation to the infratentorial enlargement more neurological deficits occur [37]. Hunter et al. stated that the vast majority (88%) of these tumors grow [38]. According to both publications, active treatment seems to be necessary, at least as soon as tumor growth is observed.

Besides the advantages of complex skull base approaches, the workhorse for tumors of the cerebello-pontine angle is the retrosigmoid suboccipital approach, which can be combined with a tailored supratentorial approach in case of extension to the middle and anterior fossa [10, 12,13,14, 39]. These approaches were mainly used in our series.

The overall results of the whole study group of 133 primary PCMs are 27.3% new CN deficits and 59.4% GTR, which is in the range of major published series, in which different skull base approaches have been used [2,3,4,5,6,7,8,9,9,10,11,12,13,14, 22,23,24,25,26,27,28,29,30,31]. CN deficits are described in 22–67% and GTR was achieved between 28 and 63%. In a recent meta-analysis, Di Carlo et al. evaluated twelve studies with respect to CN deficits, dichotomized to the retrosigmoid and to petrosal approaches [40]. They found that CN7 palsy was more often observed in the RSA (16.6% vs 11.4%), whereas CN4 palsy occurred at a higher rate in the petrosal approach (7.6% vs 2.1%). This is consistent with our observation that CN7 and CN 8 have a relatively high risk for new deficits (8 and 11%) during PCM resection via the RSA, which is explained by their location within the surgical work path. Although the meta-analysis included only studies with a relatively low number of patients, the results are important for the preoperative risk stratification in regard to different approaches. Altogether, permanent CN deficits occurred in these studies in a low range and were superior to those data presented in larger series.

Our data include 25 patients with recurrent PCMs. Operating on these lesions has a restriction in itself. If the primary tumor was not completely resectable during the first surgery, one cannot expect to be more successful in the second session. If the tumor had a malignant grade, the risk for fast regrowth was high. Therefore, the indication for repeated surgery has to be critically evaluated and all were done on an individual basis. We saw an indication when significant regrowth of tumor remnants associated with brainstem compression was seen after the first surgery. The tumors were already irradiated and regrew. The tumors were WHO grade 2 and 3 meningiomas, had been already irradiated and no alternative treatment options were left. Most of the patients with regrowing PCMs suffer already from neurological deficits, as seen in our cohort. The risk for further deficits is therefore comparably low. As shown by Li in this special group, GTR was possible in eight out of 23 patients (34.8%) [41], and surgical morbidity occurred in 36.4%. GTR in our series was achieved in only three patients (12.5%) and the morbidity rate was 20%.

Our general strategy for any lesion in the brain is to stay primarily with a standard approach, tailored in size to each lesion. The experience of others [8, 13, 14, 23, 26, 28, 40] and our own showed us that four main factors influenced the resectability of the PCMs and that these factors also influenced the outcome of the patients (1) the localization with the dural adhesion zone, (2) the consistency of the tumor, (3) the surface of the tumor, and (4) the arachnoid cleavage plane and the lack of brainstem edema. According to our results, the leading intraoperative factor was tumor consistency. Little et al. (2005) described already tumor consistency as an important factor for resectability [28] by pointing out that fibrous tumors have an increased risk because of the need for sharp dissection and tumor manipulation during debulking. In this meningioma type, cranial nerve deficits occurred in his series in 41%, which was much higher than in all others. Kawase raised the question in his comment about the possibility to diagnose relevant tumor-specific properties prior to surgery [28].

Influenced by these data, we propose a classification system for the extent of resection and CN outcome, by combining the aspects of anatomical localization and specific tumor features, as mentioned above. We focused our classification-based analysis primarily on standard approaches such as the RSA and the subtemporal/pterional craniotomy. By stratifying our results to the proposed classification system, it became obvious that significantly better results of CN preservation could be achieved in type I and type II tumor patients, in which new cranial nerve deficits occurred in only 11.9% resp. 10.0% which was significantly better in comparison to the groups III and IV with 44.4 and 43.8% (p = 0.010). Notably, GTR was achieved in 78.7% of type 1 tumors, which was significantly higher than in the types II-IV (p < 0.001). And again, the most important factor for a successful and uneventful resection was the consistency of the tumor with an arachnoid cleavage plane to the brainstem. The vascularity did not play a significant role in the preservation of CN functions, as anticipated. Therefore, we did not include vascularity as a further factor in our final classification system. Preoperative embolization of highly vascularized lesions was also not performed by the neuroradiologists in our series due to possible embolization-associated complications [42].

The ideal goal for the further development of the intraoperative classification system, as used, is the determination of a patient’s risk of undergoing surgery prior to the procedure. Therefore, tumor-specific features should be definable on preoperative MRI. In recent years MRI parameters have been identified to group meningiomas that are easier or less resectable. Pirayesh et al. reported about an excellent correlation between tumor resectability and preoperative presence or absence of an arachnoid cleavage plane and an irregular tumor margin towards the brainstem [20]. Nicosia et al. pointed out that a high tumor intensity on a T2-weighted image correlates with soft tumor consistency and vascularity, whereas brainstem edema is significantly related to surgical morbidity because of invasive growth [20, 21]. From these findings, we assessed our data with respect to preoperative MRI scans. Our special focus was on the consistency of the tumors. Preoperative MR images of 70 patients were available for evaluation. Consistent with prior reports, tumors with a hyperintense signal in T2 were highly likely to have a soft texture with a sensitivity of 95.6% and a negative predictive value of 88.2%. Our results corroborated the previous findings and corresponded well with the proposed features. Consequently, to build on these results, we have initialized a prospective study with systematic data collection on a standard protocol that allows us to use the radiological information for better planning and risk stratification in PCMs.

Over time, the knowledge about the PCM-specific properties changed our surgical understanding. We tried to achieve GTR in type I and, if possible, in type II PCMs and went for mass reduction in type III and type IV tumors as much as possible without taking risks for the quality of life of the patients, a strategy that has been also advocated by Seifert in his landmark paper [13]. Especially in type IV tumors, which are more clival meningiomas than PCMs the experience taught us that these tumors are of high risk for neurological deficits and can barely be removed via standard approaches.

Major complications in our series were new permanent hemipareses in three cases, which were in two cases associated with dysphagia that urged us to tracheotomize the patients. Severe complications occurred predominantly in patients with firm tumors or aggressive meningiomas grade 2 and 3. With a complication rate of 16.5% in primary and 20% for recurrent PCM surgery, we are in the middle of the overall common complication or morbidity rate of 11 – 45%, described in the literature [1, 13, 25, 27, 30, 33, 43, 44]. The mortality rate of PCM surgery is low and ranges between 0 and 10% [1, 13, 25, 27, 30, 33, 43, 44]. Our mortality rate of the whole group was 1.9% and occurred only in primary PCMs.

Limitations

The presented study is a retrospective analysis. Therefore, not all data were available and follow-ups were missing. Regarding the outcome analysis, we present our experience with the main focus on the retrosigmoid and the subtemporal/pterional approach and specific tumor features, implemented in an intraoperative classification system. The logical consequence of this approach is a preoperative reliable risk stratification that allows the determination of specific tumor characteristics by MRI prior to surgery. A comparison between standard approaches and skull base approaches with a special focus on tumor-specific properties is pending to select the appropriate surgical approach on an individual basis.

留言 (0)

沒有登入
gif