External validation of current prediction systems of improvement after decompression surgery in Chiari malformation type I patients: can we do better?

Use of Artificial Intelligence in the Prediction of Chiari Malformation Type 1 Recurrence After Posterior Fossa Decompressive Surgery

Purpose The purpose of this study was to train a deep learning-based method for the prediction of postoperative recurrence of symptoms in Chiari malformation type 1 (CM1) patients undergoing surgery. Studies suggest that certain radiological and clinical features do exist in patients with treatment failure, though these are inconsistent and poorly defined. Methodology This study was a retrospective cohort study of patients who underwent primary surgical intervention for CM1 from January 2010 to May 2020. Only patients who completed pre- and postoperative 12-item short form (SF-12) surveys were included and these were used to classify the recurrence or persistence of symptoms. Forty patients had an improvement in overall symptoms while 17 had recurrence or persistence. After magnetic resonance imaging (MRI) data augmentation, a ResNet50, pre-trained on the ImageNet dataset, was used for feature extraction, and then clustering-constrained attention multiple instance learning (CLAM), a weakly supervised multi-instance learning framework, was trained for prediction of recurrence. Five-fold cross-validation was used for the development of MRI only, clinical features only, and a combined machine learning model. Results This study included 57 patients who underwent CM1 decompression. The recurrence rate was 30%. The combined model incorporating MRI, pre-operative SF-12 physical component scale (PCS), and extent of cerebellar ectopia performed best with an area under the curve (AUC) of 0.71 and an F1 score of 0.74. Conclusion This is the first study to our knowledge to explore the prediction of postoperative recurrence of symptoms in CM1 patients using machine learning methods and represents the first step toward developing a clinically useful prognostication machine learning model. Further studies utilizing a similar deep learning approach with a larger sample size are needed to improve the performance.

A Novel Craniocervical Junction Compression Severity Index-Based Grading System for Multidirectional Quantification of the Biomechanics at Foramen Magnum of Chiari Malformation Type I

Objective  This study aimed to establish a novel grading system, based on the craniovertebral junction compression severity index (CVJCSI) for multidirectional quantification at the foramen magnum plane for Chiari malformation type I (CMI). Methods  The CVJCSI grading system was established to stratify patients based on the ventral (modified clivoaxial angle < 138°), dorsal (tonsil herniation), and central (brainstem herniation) CVJ (craniovertebral junction) compression, the CVJCSI grading system was established to stratify patients. The optimal surgical method for each grade was recommended by intragroup comparisons regarding the efficacy of the three operations. Finally, according to the CVJCSI grading system, a prospective validation trial was performed and surgically treated for internal validation. Results  Based on the retrospective study ( n  = 310), the CVJCSI included six grades: I: syrinx alone without compression; II: dorsal compression; III: dorsal and central compression; IV: ventral compression; V: dorsal and ventral compression; and VI: ventral, dorsal, and central compression. Among all available variables, only the CVJCSI and surgical methods significantly affected the CCOS. The CCOS scores, overall and for each CVJCSI grade, increased in the prospective cohort ( n  = 42) compared with that in the retrospective analysis. Conclusions  The CVJCSI can be used to stratify CMI patients. The higher the CVJCSI grade, the more severe the CVJ compression and the worse posterior fossa deformity. Meanwhile, the CVJCSI was negatively correlated with the CCOS. The lower the CVJCSI grade, the better the response to surgery, and the less-invasive surgical procedures were warranted. Finally, the prospective cohort study validated the proposed CVJCSI-based surgical protocols.

Individualized surgical treatment of Chiari 1 malformation: A single-center experience

OBJECTIVE

To present our experience with an individualized surgical approach to treat Chiari malformation type 1.

METHODS

Based on (1) neurological symptoms, (2) the existence and extent of a syrinx and (3) the degree of the tonsillar descent we performed four types of approaches on a case-by-case basis in 81 patients: (1) foramen magnum decompression (FMD) with dura splitting (FMDds); (2) FMD with duraplasty (FMDdp); (3) FMD with duraplasty and tonsillar manipulation (FMDao); and (4) tonsillar resection/reduction (TR). Patient characteristics, Chiari Severity Index (CSI), fourth ventricular roof angle (FVRA) and Chicago Chiari Outcome Scale (CCOS) were analyzed.

RESULTS

CCOS was between 13 and 16 points in 8/11 (73 %) patients after FMDds, 38/45 (84 %) patients after FMDdp, and 24/24 (100 %, one patient lost to follow-up) patients after TR. We experienced an overall complication rate of 13.6 % (11/81) in this series, whereas seven of these eleven complications (64 %) occurred in the FMDao group and the complication rate increased with the invasiveness of the approach (0 % FMDds; 4 % FMDdp; 12 % TR).

CONCLUSION

Given the clear correlation between the extend of the approach and the complication rate the least invasive approach necessary to achieve clinical improvement should be selected. Due to the high complication rates, FMDao should not be used as a treatment option. The severity of the tonsillar descent, basilar invagination and current CM1 scores could be used to aid in the approach selection.

Chiari Type 1 Malformation and Syringomyelia in Children: Classification and Treatment Options

Chiari type 1 malformation (CIM) is defined as tonsillar ectopia of >5 mm, while syringomyelia (SM) is defined as a cerebrospinal fluid (CSF)-filled cavity larger than 3 mm dissecting the spinal cord. Over the last decades, our understanding of these pathologies has grown; however, many controversies still exist almost in every aspect of CIM and SM, including etiology, indication for treatment, timing of treatment, surgical technique, follow-up regime, and outcome. This chapter provides a comprehensive overview on different aspects of CIM and SM and on the still existing controversies, based on the evidence presently available. Future directions for clinical research concerning CIM and SM treatment and outcome are elaborated and discussed as well.

Relationship Between Syrinx Resolution and Cervical Sagittal Realignment Following Decompression Surgery for Chiari I Malformation Related Syringomyelia Based on Configuration Phenotypes

OBJECTIVE

Combined with different configuration types of syringomyelia, to analyze the correlation between syrinx resolution and changes in cervical sagittal alignment following Foramen magnum and Magendie dredging (FMMD) for syringomyelia associated with Chiari I malformation (CM-I), and to further explore the respective relationship with clinical outcome.

METHODS

A consecutive series of 127 patients with CM-I and syringomyelia who underwent FMMD in our center met the inclusion criteria of this study. Their clinical records and radiologic data were retrospectively reviewed. The Japanese Orthopedic Association (JOA) scoring system and the Chicago Chiari Outcome Scale (CCOS) were used to evaluate the surgical efficacy. The phenotypes of syringomyelia and the clinical characteristics of the patients were analyzed according to grouping by cervical curvature at baseline.

RESULTS

The preoperative straight or kyphotic cervical alignment is more common in the moniliform syrinx. After surgery, the syrinx resolution and cervical sagittal realignment in the moniliform group are more obvious, and the corresponding prognosis is relatively better. Spearman correlation analysis showed that the ΔS/C ratio (the change ratio of syrinx/cord) was positively correlated with the CCOS (p = 0.001, r = 0.897) and ΔC2-7A (the change of lower cervical angle) (p = 0.002, r = 0.560). There was also a correlation between the ΔJOA score (the change rate of the JOA score) and ΔC2-7A (p = 0.012, r = 0.467).

CONCLUSION

After decompression surgery, syrinx resolution may coexist with the changes in the subaxial lordosis angle, especially for syrinx in moniliform type, and the relationship between syrinx resolution and cervical sagittal realignment might be valuable for evaluating the surgical outcome.

Dura Management Strategies in the Surgical Treatment of Adult Chiari Type I Malformation: A Retrospective, Multicenter, Population-Based Parallel Cohort Case Series

BACKGROUND

Symptomatic Chiari I malformation is treated with suboccipital decompression and C1 laminectomy. However, whether the dura should be opened (durotomy) or enlarged with a graft (duraplasty) remains unclear.

OBJECTIVE

To compare outcomes in adult Chiari I malformation patients treated with duraplasty, durotomy, or without dural opening ("mini-decompression").

METHODS

A retrospective, multicenter, population-based cohort study was performed of all adult patients surgically treated for a Chiari I malformation at 3 regional neurosurgical centers between 2005 and 2017. Three different dura management strategies were favored by the participating hospitals, with data stratified accordingly. The primary outcome was measured using the Chicago Chiari Outcome Scale (CCOS), dichotomized into favorable (CCOS ≥13) or unfavorable (CCOS ≤12). Propensity score matching was used to adjust for potential confounders in outcome comparisons.

RESULTS

In total, 318 patients were included, of whom 52% were treated with duraplasty, 37% with durotomy, and 11% with mini-decompression. In total, 285 (90%) showed a favorable surgical outcome (CCOS ≥13). Duraplasty was associated with more favorable CCOS and shorter hospital stay compared with durotomy, both in unadjusted (93% vs 84%. P = .018 and 6.0 vs 8.0 days, P < .001) and adjusted analyses (92% vs 84%, P = .044 and 6.0 vs 8.0 days, P < .001). Mini-decompression was excluded from the adjusted analyses because of its small sample size.

CONCLUSION

In this study of adult Chiari I malformation, posterior fossa decompression with duraplasty was associated with more favorable postoperative outcome, as determined by the CCOS, compared with posterior fossa decompression with durotomy alone.

Incorporating patient-centered quality-of-life measures for outcome assessment after Chiari malformation type I decompression in a pediatric population: a pilot study

OBJECTIVE

Optimal management of pediatric Chiari malformation type I (CM-I) is much debated, chiefly due to the lack of validated tools for outcome assessment, with very few tools incorporating patient-centered measures of health-related quality of life (HRQOL). Although posterior fossa decompression (PFD) benefits a subset of patients, prediction of its impact across patients is challenging. The primary aim of this study was to investigate the role of patient-centered HRQOL measures in the assessment and prediction of outcomes after PFD.

METHODS

The authors collected HRQOL data from a cohort of 20 pediatric CM-I patients before and after PFD. The surveys included assessments of selected Patient-Reported Outcomes Measurement Information System (PROMIS) health domains and were used to generate the PROMIS preference (PROPr) score, which is a measure of HRQOL. PROMIS is a reliable standardized measure of HRQOL domains such as pain, fatigue, depression, and physical function, which are all relevant to CM-I. The authors then compared the PROPr scores with Chicago Chiari Outcome Scale (CCOS) scores derived from time-matched clinical documentation. Finally, the authors used the PROPr scores as an outcome measure to predict postsurgical HRQOL improvement at 1 year on the basis of patient demographic characteristics, comorbidities, and radiological and physical findings. The Wilcoxon signed-rank test, Mann-Whitney U-test, and Kendall's correlation were used for statistical analysis.

RESULTS

Aggregate analysis revealed improvement of pain severity after PFD (p = 0.007) in anatomical patterns characteristic of CM-I. Most PROMIS domain scores trended toward improvement after surgery, with anxiety and pain interference reaching statistical significance (p < 0.002 and p < 0.03, respectively). PROPr scores also significantly improved after PFD (p < 0.008). Of the baseline patient characteristics, preexisting scoliosis was the most accurate negative predictor of HRQOL improvement after PFD (median -0.095 vs 0.106, p < 0.001). A correlation with modest magnitude (Kendall's tau range 0.19-0.47) was detected between the patient-centered measures and CCOS score.

CONCLUSIONS

The authors observed moderate improvement of HRQOL, when measured using a modified panel of PROMIS question banks, in this pilot cohort of pediatric CM-I patients after PFD. Further investigations are necessary to validate this tool for children with CM-I and to determine whether these scores correlate with clinical and radiographic findings.