Development and internal validation of nomogram for chronic subdural hematoma recurrence after surgery

Introduction

Chronic subdural hematoma (CSDH) is a prevalent condition among the elderly. Population-based studies show that the yearly incidence of CSDH varies between 8.2 and 17.6 cases per 100,000 individuals (1). The prevalence of CSDH has grown as a result of population aging and increased usage of antiplatelet and anticoagulant drugs (2). The treatment of choice for symptomatic patients is surgery, which includes craniotomies, twist drills, and burr-holes (3,4). Recurrence rates after surgery ranged from 10.1% to 29% in previous studies (5-8).

Nomogram is a two-dimensional (2D) calculator that uses predictors and a mathematical function to forecast a range of medical conditions (9,10). According to the literature review, the following are risk factors for CSDH recurrence: age, seizure, coagulopathy, the type of CSDH, the thickness of the hematoma, postoperative midline shift, postoperative subdural air collection, and postoperative re-expansion. Nevertheless, there is ongoing debate regarding recurring risk factors, and no consensus has been achieved (11-17). In the studies of Chon et al. and Ko et al. (11,12), antiplatelet and anticoagulant medications were strongly associated with CSDH recurrence, whereas earlier researches showed that these drugs were not an independent predictor (13-15).

Prior studies revealed that mixed-density subtype an is commonly considered at higher risk for recurrence (16,17); however, Hammer et al. discovered no link between CSDH density and recurrence (13). Additionally, bilateral CSDHs have been associated with higher recurrence rates in previous studies (14); however, previous retrospective study shows no significant independent association (15). Because the recurrence of CSDH is a potential complication during follow-up, the time-to-event analysis may be preferred over binary logistic regression. Buakate et al. used prognosis Cox hazard regression to predict CSDH recurrence, and found that clopidogrel, operation time 90 minutes or more, amount of saline irrigation 3,200 mL or more were all significantly associated with recurrence of CSDH (18). Moreover, the review of literature revealed a lack of evidence on nomograms, a scoring-based method for predicting CSDH recurrence (17,18).

The primary objective was to explore the predictors of CSDH recurrence after surgery. Additionally, a secondary objective was to create a clinical nomogram for predicting CSDH recurrence using the time-to-event analysis technique. We present this article in accordance with the STROBE reporting checklist (available at https://jxym.amegroups.com/article/view/10.21037/jxym-24-83/rc).

Methods

Study design and study population

A retrospective study was conducted on all CSDH patients who underwent surgery from January 2009 to December 2022. CSDH was diagnosed using cranial computed tomography (CT), which showed subdural collections that were consistent with chronic hematoma. Moreover, intraoperative confirmation was required, with operative notes documenting the presence of old blood and/or hematoma membranes. Therefore, we excluded CSDH patients with no preoperative CT, no postoperative cranial CT, and no medical records, and previous surgery from CSDH. Moreover, individuals with bilateral CSDH were excluded due to the challenges in assessing re-expansion post-surgery and evaluating recurrence in the presence of discordant findings.

One hundred forty-one patients were recruited from a prior study conducted by Buakate and Tunthanathip (18), while an additional 35 patients were included in the study. Therefore, seven patients were thus excluded in accordance with the exclusion criteria. In detail, two patients were not available for preoperative CT scans in the database, and five people had bilateral CSDHs. Consequently, the present study examined 169 CSDH patients, as illustrated in Figure 1.

Figure 1 Patient selection flowchart. CSDH, chronic subdural hematoma; CT, computed tomography.

Clinical features and imaging results, such as age, gender, cause of CSDH that based on medical record, Glasgow Coma Scale (GCS) scores, signs, symptoms, underlying disease, and preoperative medication, were gathered from patients on a structured record form. The preoperative GCS scores were assessed on the date of admission and were divided into three groups based on severity: GCS scores 13–15, GCS scores 9–12, and GCS levels 3–8 (19). Nakaguchi et al. classified the kind of CSDH comprised homogeneous, laminar, separated, gradation, and trabecular subtypes (20). The type of operation, drain position, number of burr holes, operation duration, and drain placement were all evaluated. Additionally, the computerized database contained the amount of irrigated normal saline and the surgical time (the interval between the incision and full wound closure) for each case. Based on Mori and Maeda’s study (17), the percentage of cortical re-expansion following surgery was assessed from preoperative and postoperative CT scans. Furthermore, postoperative pneumocephalus was reviewed according to the research conducted by Ihab (21).

The recurrence of CSDH was characterized by either a deterioration in the patient’s neurological condition during follow-up or an increase in the maximal thickness of CSDH on the surgically treated side to 10 mm or greater on the follow-up CT scans (22). Cranial CT scans were utilized to examine the recurrent occurrence at every post-operative visit until March 31, 2023. Moreover, two independent neurosurgeons were evaluated preoperative and postoperative CT scans. When the reviewers disagreed, they discussed it and came to a conclusion.

Procedure

All patients underwent operation with general anesthesia; therefore, patients were placed on a headrest in a supine position. Type of operation, surgical technique and clinical management were selected based on patient’s status, imaging findings and surgeon’s preference. Due to the context of the present study being a tertiary hospital, certain patients were performed by the chief training resident under supervision.

For burr hole craniostomy, the most cases of burr hole craniostomy with two burr-hole was performed at frontal and parietal areas. Ten milliliters of 1% xylocaine and 1:200,000 adrenaline were used to infiltrate the incision site. The dura mater was first coagulated before being opened with a cruciate incision. The subdural bloody fluid was then flushed with room-temperature saline until the solution became clear. Drains were employed based on the surgeon’s preference. The subgaleal drain was positioned under the galea aponeurotica, while the subdural drain was placed beneath the dura mater. Drain connected to a soft collection bag that was suspended on the side of the bed and was withdrawn approximately 48 hours after the operation.

For craniotomy, the decision between a small or large craniotomy procedure depended on the surgeon’s judgment. After dura opening, partial membranectomy was performed; then, normal saline was irrigated until the surgical filed was clear. The kind of drain was selected depending on the surgeon’s decision. While the subgaleal drain was positioned under galea aponeurotica, the subdural drain was inserted under the dura matter. The subdural drain was then connected to the collection bag without applying negative pressure, whereas the subgaleal drain was linked to the collection bag using negative pressure. Additionally, the operative time, defined as the duration from skin incision to skin closure, and the volume of normal saline used for intraoperative irrigation were documented in the electronic operative note.

A postoperative CT scan was generally done within 24 hours following the operation. Further CT scans during the same hospitalization were requested based on the surgeon’s judgment. Every appointment at the outpatient clinic, a CT scan of the brain was done until the CSDH was completely resolved.

Statistical analysis

Descriptive statistics were utilized to determine clinical characteristics and imaging results. Categorical data was presented as percentages, while continuous variables were presented as the mean with standard deviation (SD). The missing data was managed using complete case analysis. The predictors of both the recurrent and non-recurrent groups were assessed using a time-to-event analysis using the Kaplan-Meier curve and the log-rank test. Furthermore, the maximum selected rank statistical technique was utilized to determine the cutoff threshold for each continuous variable subject to dichotomization, such as normal saline solution volume and operative time (23).

Cox regression analysis was employed to report the factors associated with recurrence as a hazard ratio (HR) with 95% confidence interval (CI). A multivariable model with backward stepwise selection was utilized to investigate the candidate variables that had a P of 0.1 or less in the univariate analysis. The models were selected using the Akaike information criterion (AIC) with the clinical application, and P<0.05 were considered statistically significant (23). Additionally, a Schoenfeld residual test was done to evaluate the proportional hazard assumption, and the P of the test is greater than 0.05, indicating that the assumption was not violated (23,24). The statistical analysis was carried out using the R software version 4.4.0 (R Foundation, Vienna, Austria).

Nomogram development and internal validation

The multivariable analysis’s significant variables which influence CSDH recurrence were used to construct the prediction model. Based on Zhang and Kattan (24), a nomogram was created, and Harrell’s concordance index (C-index) was used to construct a discrimination measure for survival models. The bias-corrected C-index was computed for internal validation using 10-fold cross-validation and the bootstrap method 1,000 times (25). Moreover, the calibration plot was created to show how accurately the predicted probabilities were calibrated. The R program with the “regplot” package was used to carry out the 2D nomogram development and digital nomogram was created by “DynNom” package (26,27).

Ethical considerations

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The human research ethics committee of Faculty of Medicine, Prince of Songkla University approved the present study (REC 64-395-10-1). The informed consent of the patients was not necessary for the present study because it was a retrospective analysis. However, patient identification numbers were encoded before analysis.

Results

Clinical features and imaging results

The clinical characteristics of the 169 CSDH patients in the present study are displayed in Table 1. The patient’s average age was 61.17 years (SD =16.88 years), and 79.3% of them were male. The most prevalent underlying disease in the current cohort was hypertension, and 16.8% had diabetes mellitus. Preoperative aspirin usage was found in 21.3%, while 8.3% of the present cohort had clopidogrel usage before the operation. Furthermore, 5.9% of patients had preoperative warfarin use. Seventy-four percent of the total cases had a history of traumatic brain injury, whereas 23% of CSDH cases had an unknown etiology. GCS 13–15, GCS 9–12, and GCS 3–8 were noted before surgery in 86.4%, 5.9%, and 7.7%, respectively.

Preoperative cranial CT findings are revealed in Table 2. The homogenous hypointensity type was the most common type of CSDH, while the laminar type was found in 17.8%. The frontoparietal convexity was the location of nearly all of the CSDH. Additionally, the preoperative midline shift of the CSDH was 8.67 mm, whereas the preoperative mean width of the CSDH was 20.53 mm (SD =7.44 mm). In 69.8% of instances, the basal cistern had been obliterated before surgery.

Table 3 shows the treatment and result of surgery. Burr holes with irrigation accounted for 89.9% of operations, while craniotomies were carried out in 10.1% of all instances. The average operative duration was 92.20 minutes (SD =46.98 minutes), and 56.8% of cases had obvious postoperative intracranial air. The average follow-up time was 160.30 days (SD =196.45 days), and the recurrence incidence was 22.5% of total cases. During follow-up, no one died in the present study.

Factors associated with post-operative CSDH recurrence

Median follow-up time was 99 days [interquartile range (IQR), 96 days] and the recurrence rate of CSDH was found to be 22.5%. Moreover, median time of recurrence was 75 days (IQR, 31 days). Using Cox hazard regression, a univariate analysis was conducted on various clinical variables. Initially, the significant factors were preoperative antiplatelet use (HR =2.27, 95% CI: 1.13–4.54), preoperative GCS scores (GCS 13–15 = reference; HR of GCS 9–12 =3.17, 95% CI: 1.10–9.15; and HR of GCS 3–8 =4.23, 95% CI: 1.82–9.83), operative time 140 minutes or more (HR =0.22, 95% CI: 0.09–0.55), and amount of saline irrigation 3,200 mL or more (HR =0.41, 95% CI: 0.02–0.85), respectively. Moreover, the Kaplan-Meier curves of each predictor are shown in Figure 2A-2E.

Figure 2 Kaplan-Meier curve. (A) Overall; (B) antiplatelet; (C) GCS score; (D) Optime; (E) amount of saline irrigation. GCS, Glasgow Coma Scale; NSS, normal saline solution; Optime, operative time.

As a result, a backward stepwise selection technique was used to evaluate the potential variables associated with CSDH recurrence. Preoperative GCS scores (GCS 13–15 = reference; HR of GCS 9–12 =3.07, 95% CI: 1.05–8.92, P=0.03; and HR of GCS 3–8 =2.84, 95% CI: 1.21–6.66, P=0.01), preoperative antiplatelet use (HR =2.97, 95% CI: 1.46–6.02, P=0.002), operative time (HR =0.30, 95% CI: 0.12–0.74, P=0.009), and the amount of saline irrigation (HR =0.05, 95% CI: 0.007–0.74, P=0.003) were predictors in the final model with the lowest AIC, as shown in Table 4.

The 2D nomogram was developed using the significant parameters in the multivariable analysis as shown in Figure 3. Therefore, Harrell’s C-index was 0.822 for the discrimination measure of the predictive model. For internal validation, the bias-corrected C-indexes of bootstrap and 10-fold cross-validation methods were 0.821 and 0.821 respectively. Additionally, all P of the global Schoenfeld test for each variable were more than 0.05, indicating that the Cox hazard regression assumption was not violated, as shown in Figure 4A. Moreover, the calibration plot was near a 45-degree line, as shown in Figure 4B.

Figure 3 The nomogram predicts 30-, 60-,90-,120- and 150-day survival probabilities. To use the nomogram, draw a straight line upward from the patient’s characteristics of the GCS, antiplatelet, Optime, and the amount of irrigated saline. Then, the sums of the scores of all variables on the total points scale, draw another straight line down from the scale of the total points through the 30-, 60-, 90-, 120- and 150-day survival probabilities of recurrence. FU, follow-up; GCS, Glasgow Coma Scale; NSS, normal saline solution; Optime, operative time.

Figure 4 Assumption test of Cox regression and goodness-of-fit test of nomogram. (A) Global Schoenfeld test each predictor. (B) Calibration plot. GCS, Glasgow Coma Scale; NSS, normal saline solution; Optime, operative time.

Additionally, a 2D nomogram was changed to digital nomogram for user-friendly visual format and creating personalized survival curve, as shown in Figure 5. The R script for producing digital nomogram was proposed via https://github.com/thara7640/DynNom_CSDH.

Figure 5 The screenshot of nomogram program for predicting CSDH recurrence. CSDH, chronic subdural hematoma; FU, follow-up; GCS, Glasgow Coma Scale; NSS, normal saline solution; Optime, operative time.

Discussion

Following surgery, the natural course of CSDH was either resolved or recurrent. In this investigation, the recurrence rate of CSDH was found to be 22.5%. This is consistent with earlier studies that indicated a recurrence risk of CSDH following surgery ranging from 10.1% to 29% (5-8). Furthermore, previous research employed logistic regression analysis, whereas the current study used time-to-event analysis, which is better suited to dealing with different time endpoints for recurring criteria spanning from 3 to 6 months (28-30).

As the result, predictors significantly associated with CSDH recurrence after surgery in the present study were preoperative antiplatelet use, preoperative GCS scores, operative time, and amount of saline irrigation. The use of antithrombotic drugs, such as antiplatelets or anticoagulants, has increased as a result of the patient’s underlying illness. Approximately 35% of CSDH patients in the present study used antithrombotic medications prior to surgery, and antiplatelet was significantly related to recurrence. This could be explained by variations in the resuming time of antiplatelet medicines in the postoperative period (31). An early restart of antithrombotic medicine may result in a recurrence of CSDH; nevertheless; however, data from the present study did not include precise records on when patients restarted these medications. Hence, the best resumption period should be studied in the future. According to previous research, intraoperative saline irrigation significantly reduces CSDH recurrence when compared to no irrigation (32,33). Therefore, the relationship between irrigation volume and recurrence remains a subject of debate. The results could be explained by irrigation washing out many inflammatory cytokines and fibrinolytic substances; this strategy may assist prevent postoperative recurrence (34). Surgical time was related with a low recurrent incidence, which could be attributed to the volume of saline irrigation, or the extended duration of operation produced by the training residents’ performance. One limitation of the present study is that we did not routinely record the temperature of the irrigation fluid utilized during the procedure. The temperature of the fluid could affect hemostasis since warmer irrigation could perhaps induce coagulation and lower microbleeding while colder irrigation could have the reverse effect (35). While our standard practice involves using room-temperature saline, variations in temperature across cases could not be assessed in our analysis. Future studies should consider standardizing and recording irrigation temperature to determine its potential impact on recurrence rates.

A nomogram is a clinical prediction tool that has been examined in numerous fields of medicine, including neuro-oncology and traumatic brain injury (9,10). Previous research has established a nomogram based on binary logistic regression for predicting the recurrence of CSDH (36). Nomogram built from Cox regression has benefits because it provides multi-factorial risk estimation, personalized prognosis, and a user-friendly visual format (37). The nomogram functioned as a 2D visual scoring system that relied on visual assessment, which led to inconsistent scoring (38). An alternative way of assigning consistency scores should be provided, such as computer software that may generate personalized survival curves in a user-friendly visual design. To the best of the authors’ knowledge, this is the first study that demonstrated the clinical nomogram was effective in predicting CSDH recurrence. However, there are certain limitations to the research that should be addressed.

Because study population were collected from long period, some practice of CSDH treatment has been changed, such as non-invasive intervention. Middle meningeal artery (MMA) embolization has been proposed as the alternative methods to treat CSDH (39,40). Davies et al. conducted a prospective, multicenter trial that assessed adjunctive MMA embolization following surgery. Their findings indicated that MMA embolization in conjunction with surgery was associated with a reduced risk of hematoma recurrence or progression (40). In Prince of Songkla University, this intervention has recently been performed in CSDH cases since 2023; therefore, surgical operation was treatment of choice in the present study. The nomogram from the current study has to be validated with new cases. As a result, external validation would be the next step in evaluating the prediction tool’s performance on previously unseen data from novel patients at the same centers where the nomogram was developed (temporal validation), or patients from centers that were not involved in the tool’s development (geographical validation) (41). In addition, because the retrospective study design may introduce bias, prospective multicenter research should be conducted in the future to validate nomogram’s performance (41,42).

Conclusions

In summary, clinical variables significantly associated with CSDH recurrence were used to build clinical prediction tools that could aid physicians’ decision-making and care of high-risk patients in general practice.

Acknowledgments

The authors would like to offer their special thanks to Professor Nakornchai Phuenpathom, and Associate professor. Sakchai Sae-heng for their advice about manuscript preparation. This research was part of a study project that will be published elsewhere, while this study focused on nomogram for predicting CSDH recurrence.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jxym.amegroups.com/article/view/10.21037/jxym-24-83/rc

Data Sharing Statement: Available at https://jxym.amegroups.com/article/view/10.21037/jxym-24-83/dss

Peer Review File: Available at https://jxym.amegroups.com/article/view/10.21037/jxym-24-83/prf

Funding: None.

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at https://jxym.amegroups.com/article/view/10.21037/jxym-24-83/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The human research ethics committee of Faculty of Medicine, Prince of Songkla University approved the present study (REC 64-395-10-1). The informed consent of the patients was not necessary for the present study because it was a retrospective analysis. However, patient identification numbers were encoded before analysis.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Rauhala M, Luoto TM, Huhtala H, et al. The incidence of chronic subdural hematomas from 1990 to 2015 in a defined Finnish population. J Neurosurg 2019;132:1147-57. [Crossref] [PubMed]
2. Miah IP, Holl DC, Peul WC, et al. Dexamethasone therapy versus surgery for chronic subdural haematoma (DECSA trial): study protocol for a randomised controlled trial. Trials 2018;19:575. [Crossref] [PubMed]
3. Ryu SM, Yeon JY, Kong DS, et al. Risk of Recurrent Chronic Subdural Hematoma Associated with Early Warfarin Resumption: A Matched Cohort Study. World Neurosurg 2018;120:e855-62. [Crossref] [PubMed]
4. Weigel R, Schmiedek P, Krauss JK. Outcome of contemporary surgery for chronic subdural haematoma: evidence based review. J Neurol Neurosurg Psychiatry 2003;74:937-43. [Crossref] [PubMed]
5. Ohba S, Kinoshita Y, Nakagawa T, et al. The risk factors for recurrence of chronic subdural hematoma. Neurosurg Rev 2013;36:145-9; discussion 149-50. [Crossref] [PubMed]
6. Oh HJ, Lee KS, Shim JJ, et al. Postoperative course and recurrence of chronic subdural hematoma. J Korean Neurosurg Soc 2010;48:518-23. [Crossref] [PubMed]
7. Schwarz F, Loos F, Dünisch P, et al. Risk factors for reoperation after initial burr hole trephination in chronic subdural hematomas. Clin Neurol Neurosurg 2015;138:66-71. [Crossref] [PubMed]
8. Maroufi SF, Farahbakhsh F, Macdonald RL, et al. Risk factors for recurrence of chronic subdural hematoma after surgical evacuation: a systematic review and meta-analysis. Neurosurg Rev 2023;46:270. [Crossref] [PubMed]
9. Oearsakul T, Tunthanathip T. Development of a nomogram to predict the outcome of moderate or severe pediatric traumatic brain injury. Turk J Emerg Med 2022;22:15-22. [Crossref] [PubMed]
10. Tunthanathip T, Phuenpathom N, Jongjit A. Prognostic factors and clinical nomogram for in-hospital mortality in traumatic brain injury. Am J Emerg Med 2024;77:194-202. [Crossref] [PubMed]
11. Chon KH, Lee JM, Koh EJ, et al. Independent predictors for recurrence of chronic subdural hematoma. Acta Neurochir (Wien) 2012;154:1541-8. [Crossref] [PubMed]
12. Ko BS, Lee JK, Seo BR, et al. Clinical analysis of risk factors related to recurrent chronic subdural hematoma. J Korean Neurosurg Soc 2008;43:11-5. [Crossref] [PubMed]
13. Hammer A, Tregubow A, Kerry G, et al. Predictors for Recurrence of Chronic Subdural Hematoma. Turk Neurosurg 2017;27:756-62. [Crossref] [PubMed]
14. Torihashi K, Sadamasa N, Yoshida K, et al. Independent predictors for recurrence of chronic subdural hematoma: a review of 343 consecutive surgical cases. Neurosurgery 2008;63:1125-9; discussion 1129. [Crossref] [PubMed]
15. Lampros M, Katsiou IT, Kafritsas G, et al. Risk Factors for Recurrence in Patients Surgically Treated for Chronic Subdural Hematomas: A Single Institutional Experience. Surgeries 2025;6:19.
16. Miah IP, Tank Y, Rosendaal FR, et al. Radiological prognostic factors of chronic subdural hematoma recurrence: a systematic review and meta-analysis. Neuroradiology 2021;63:27-40. [Crossref] [PubMed]
17. Mori K, Maeda M. Surgical treatment of chronic subdural hematoma in 500 consecutive cases: clinical characteristics, surgical outcome, complications, and recurrence rate. Neurol Med Chir (Tokyo) 2001;41:371-81. [Crossref] [PubMed]
18. Buakate K, Tunthanathip T. Factors Associated with Recurrence in Chronic Subdural Hematoma following Surgery. J Health Allied SciNU 2024;14:85-93.
19. Taweesomboonyat C, Kaewborisutsakul A, Tunthanathip T, et al. Necessity of in-hospital neurological observation for mild traumatic brain injury patients with negative computed tomography brain scans. J Health Sci Med Res 2020;38:267-74.
20. Nakaguchi H, Tanishima T, Yoshimasu N. Factors in the natural history of chronic subdural hematomas that influence their postoperative recurrence. J Neurosurg 2001;95:256-62. [Crossref] [PubMed]
21. Ihab Z. Pneumocephalus after surgical evacuation of chronic subdural hematoma: Is it a serious complication? Asian J Neurosurg 2012;7:66-74. [Crossref] [PubMed]
22. Kim DH, Kim HS, Choi HJ, et al. Recurrence of the Chronic Subdural Hematoma after Burr-Hole Drainage with or without Intraoperative Saline Irrigation. Korean J Neurotrauma 2014;10:101-5. [Crossref] [PubMed]
23. Supbumrung S, Kaewborisutsakul A, Tunthanathip T. The Prognostic Value of Immunonutritional Indexes in Pineal Region Tumor. J Health Allied SciNU 2025;15:109-16.
24. Zhang Z, Kattan MW. Drawing Nomograms with R: applications to categorical outcome and survival data. Ann Transl Med 2017;5:211. [Crossref] [PubMed]
25. Antolini L, Boracchi P, Biganzoli E. A time-dependent discrimination index for survival data. Stat Med 2005;24:3927-44. [Crossref] [PubMed]
26. Marshall R. Package ‘regplot’ [Internet]. 2022. [cited 2024 Oct 15]. Available online: https://cran.r-project.org/web/packages/regplot/regplot.pdf
27. Jalali A, Roshan D, Alvarez-Iglesias A. et al. Package ‘regplot’ [Internet]. 2025. [cited 2025 Feb 15]. Available online: https://cran.r-project.org/web/packages/DynNom/DynNom.pdf
28. Santarius T, Kirkpatrick PJ, Ganesan D, et al. Use of drains versus no drains after burr-hole evacuation of chronic subdural haematoma: a randomised controlled trial. Lancet 2009;374:1067-73. [Crossref] [PubMed]
29. Leroy HA, Aboukaïs R, Reyns N, et al. Predictors of functional outcomes and recurrence of chronic subdural hematomas. J Clin Neurosci 2015;22:1895-900. [Crossref] [PubMed]
30. Liu LX, Cao XD, Ren YM, et al. Risk Factors for Recurrence of Chronic Subdural Hematoma: A Single Center Experience. World Neurosurg 2019;132:e506-13. [Crossref] [PubMed]
31. Poon MTC, Rea C, Kolias AG, et al. Influence of Antiplatelet and Anticoagulant Drug Use on Outcomes after Chronic Subdural Hematoma Drainage. J Neurotrauma 2021;38:1177-84. [Crossref] [PubMed]
32. Choi J, Whang K, Cho S, et al. Comparison of Outcomes and Recurrence in Chronic Subdural Hematoma Patients Treated by Burr-Hole Drainage with or without Irrigation. J Trauma Inj 2020;33:81-8.
33. Ishibashi A, Yokokura Y, Adachi H. A comparative study of treatments for chronic subdural hematoma: burr hole drainage versus burr hole drainage with irrigation. Kurume Med J 2011;58:35-9. [Crossref] [PubMed]
34. Jang KM, Kwon JT, Hwang SN, et al. Comparison of the Outcomes and Recurrence with Three Surgical Techniques for Chronic Subdural Hematoma: Single, Double Burr Hole, and Double Burr Hole Drainage with Irrigation. Korean J Neurotrauma 2015;11:75-80. [Crossref] [PubMed]
35. Bartley A, Bartek J Jr, Jakola AS, et al. Effect of Irrigation Fluid Temperature on Recurrence in the Evacuation of Chronic Subdural Hematoma: A Randomized Clinical Trial. JAMA Neurol 2023;80:58-63. [Crossref] [PubMed]
36. Yan C, Yang MF, Huang YW. A Reliable Nomogram Model to Predict the Recurrence of Chronic Subdural Hematoma After Burr Hole Surgery. World Neurosurg 2018;118:e356-66. [Crossref] [PubMed]
37. Tunthanathip T, Sae-Heng S, Oearsakul T, et al. Economic impact of a machine learning-based strategy for preparation of blood products in brain tumor surgery. PLoS One 2022;17:e0270916. [Crossref] [PubMed]
38. Tunthanathip T, Oearsakul T. Comparison of predicted survival curves and personalized prognosis among cox regression and machine learning approaches in glioblastoma. J Med Artif Intell 2023;6:10.
39. Debs LH, Vale FL, Walker S, et al. Middle meningeal artery embolization following surgical evacuation of symptomatic chronic subdural hematoma improves outcomes, interim results of a prospective randomized trial. J Clin Neurosci 2024;128:110783. [Crossref] [PubMed]
40. Davies JM, Knopman J, Mokin M, et al. Adjunctive Middle Meningeal Artery Embolization for Subdural Hematoma. N Engl J Med 2024;391:1890-900. [Crossref] [PubMed]
41. Supbumrung S, Kaewborisutsakul A, Tunthanathip T. Machine learning-based classification of pineal germinoma from magnetic resonance imaging. World Neurosurg X 2023;20:100231. [Crossref] [PubMed]
42. Jitchanvichai J, Tunthanathip T. Cost-effectiveness of intracranial pressure monitoring in severe traumatic brain injury in Southern Thailand. Acute Crit Care 2025;40:69-78. [Crossref] [PubMed]