Abstract
Background
Lung cancer is among the most prevalent and lethal malignancies worldwide. Non-intubated video-assisted thoracoscopic surgery (VATS) has demonstrated advantages in reducing hospital length of stay (LOS). However, clinical practice indicates that a substantial proportion of patients still experience prolonged length of stay (PLOS). Currently, no risk prediction model exists specifically for PLOS following non-intubated VATS in lung cancer patients. This study aims to analyze clinical data to identify risk factors associated with PLOS and to develop a predictive model.
Methods
A retrospective cohort study was conducted on patients undergoing non-intubated VATS lung cancer surgery between January 2024 and June 2025 at Shandong Provincial Hospital Affiliated to Shandong First Medical University. Data were collected via the Hospital Information System (HIS) and telephone follow-up electronic questionnaires. Categorical variables were analyzed using χ2 tests, and continuous variables were assessed with t-tests in univariate analyses. Variables with statistical significance in univariate analysis were entered into multivariable logistic regression to identify independent predictors and construct the prediction model. A nomogram was created for visualization. Model discrimination was assessed using the area under the receiver operating characteristic (ROC) curve, and calibration was evaluated with calibration plots.
Results
Of 742 patients analyzed, 216 had a prolonged LOS (≥8 days). PLOS was associated with significantly higher comorbidity burdens, more complex surgeries, and worse postoperative outcomes, including a greater complication rate (48.6% vs. 20.0%) than the normal LOS group (all P<0.001). Multivariable analysis identified older age [odds ratio (OR) =1.053], longer preoperative wait (OR =7.729), postoperative complications (OR =2.970), and chest tube drainage >200 mL as independent risk factors for PLOS, while body mass index (BMI) ≥30.0 kg/m2 was protective (OR =0.043). The resulting predictive nomogram demonstrated excellent discrimination with an area under the curve (AUC) of 0.943.
Conclusions
This prediction model shows robust accuracy in identifying lung cancer patients at high risk of PLOS after non-intubated VATS. It provides a theoretical basis for early identification and timely intervention by clinical staff.
Keywords: Lung cancer, video-assisted thoracoscopic surgery (VATS), non-intubated, length of hospital stay, postoperative
Highlight box.
Key findings
• The study developed a robust risk prediction model for prolonged length of stay in lung cancer patients undergoing non-intubated video-assisted thoracoscopic surgery (VATS), demonstrating high discriminative ability, good calibration, and superior clinical utility, thus providing a valuable tool for early identification and intervention in high-risk patients.
What is known and what is new?
• Non-intubated thoracoscopy, as an advanced minimally invasive technique, has demonstrated significant advantages (e.g., reduced intubation-related trauma and faster recovery) in lung cancer resection surgery. However, despite an overall reduction in hospitalization time, prolonged postoperative length of stay (LOS) remains a common clinical issue.
• This study is the first to develop a prediction model for prolonged postoperative LOS specifically in lung cancer patients undergoing non-intubated VATS.
What is the implication, and what should change now?
• Research has shown that the prolonged postoperative LOS of patients undergoing on-intubated VATS is not a random occurrence but can be predicted using a set of objective pre- and intraoperative indicators. This shift enables clinical decision-making to evolve from a model based on generalized experience to one grounded in individualized risk prediction, thereby providing a concrete tool for implementing precision accelerated recovery surgery.
• Surgeons should utilize predictive model for personalized preoperative risk assessment. For patients identified as high-risk, this enables clinicians to initiate early communication with patients and their families regarding a potentially extended recovery, collaboratively establish realistic rehabilitation expectations, and formulate proactive discharge plans, thereby enhancing patient satisfaction and engagement.
Introduction
The International Agency for Research on Cancer (IARC) of the World Health Organization reported a marked global increase in cancer incidence, with 20 million new cases recorded in 2022. Lung cancer is the most frequently diagnosed malignancy and the leading cause of cancer-related mortality worldwide, accounting for 2.5 million new cases—equivalent to 12.4% of the total global cancer burden (1). Despite the advancement in multimodal therapy tailored to disease stage, surgical resection remains a cornerstone of lung cancer management (2). With continuous innovations in medical technology and instrumentation, thoracoscopic surgery has evolved toward increasingly minimally invasive and diversified approaches. Video-assisted thoracoscopic surgery (VATS) is now widely recognized as a standard treatment modality, offering advantages over open thoracotomy such as smaller incision, faster postoperative recovery, and reduced complication rates, thereby contributing to improved patient quality of life (3).
The evolution of VATS techniques has been accompanied by continuous advancements in anesthetic strategies. Conventional VATS generally relies on general anesthesia with endotracheal intubation for perioperative pain control. However, this approach is associated with several drawbacks, including delayed recovery, sore throat, hoarseness, and potential iatrogenic injury to the airway or lung parenchyma. In response, refinements in both surgical techniques and anesthetic protocols have enabled the broader adoption of non-intubated spontaneous ventilation anesthesia for various procedures, such as lobectomy, segmentectomy, and curative lung resections. Emerging evidence suggests that non-intubated spontaneous ventilation anesthesia can expedite emergence from anesthesia, reduce pharyngeal and postoperative complications, accelerate recovery, and shorten hospital length of stay (HLOS) (4,5).
Within the current clinical framework, length of stay (LOS) has become a well-established parameter for assessing surgical quality and patient recovery. LOS functions as an indirect predictor of postoperative complication rates and the extent of recovery in cancer patients. It serves as a critical metric for evaluating perioperative care quality, reflecting the proficiency of the surgical team, adherence to standards of care, and the efficiency of medical resource utilization (6). A shorter LOS generally indicates accelerated recovery, whereas prolonged length of stay (PLOS) may be associated with delayed recovery, higher postoperative complication and mortality rates, and reduced long-term survival. In the context of rising healthcare costs and limited medical resources, LOS has garnered increasing attention from clinicians, hospital administrators, and policymakers (7). Scientifically and rationally reducing LOS after lung cancer surgery can yield substantial savings in medical resources; therefore, the development of evidence-based postoperative LOS protocols is of paramount importance. LOS is now widely recognized as a quality indicator for thoracoscopic surgery (8,9).
It is noteworthy that, despite the broad recognition of LOS as a quality indicator, significant international variations persist in its practical implementation standards. Significant international variations exist in average HLOS, reflecting disparities in healthcare systems, clinical practices, and socioeconomic factors. Countries with well-established primary care systems and integrated care pathways generally report shorter hospital stays compared to those with fragmented healthcare systems (10).
To address variability in clinical practice and promote personalized patient management, predictive tools such as nomograms provide substantial value. A nomogram is a graphical computational instrument that integrates multiple prognostic variables by translating complex multivariate regression model into an easily interpretable visual format. By employing proportionally scaled axes and alignment lines, it enables quantitative estimation of individual event probabilities based on specific predictor combinations. The clinical utility of nomograms has been extensively validated across diverse medical disciplines. In this study, we aim to develop a novel nomogram for predicting PLOS after VATS in patients with lung cancer. The purpose is to enable early identification of high-risk individuals, facilitating targeted interventions to reduce hospitalization duration, optimize clinical outcomes, enhance patient satisfaction, and improve healthcare resource allocation. Effective PLOS management, therefore, constitutes a key strategy for advancing both patient care quality and the overall efficiency of healthcare systems.
This study aims to develop a predictive model for PLOS risk by analyzing clinical data from patients undergoing non-intubated VATS for lung cancer in China. While previous international studies have investigated factors influencing PLOS after lung cancer surgery, evidence specifically addressing non-intubated VATS populations remains scarce. To fill this gap, we systematically identify independent risk factors for PLOS in this distinct surgical cohort and construct a clinically applicable predictive model. The proposed model is intended to enable early identification of high-risk patients and provide a basis for targeted interventions, thereby reducing the incidence of related complications and accelerating postoperative recovery. We present this article in accordance with the TRIPOD reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2240/rc).
Methods
Study subjects
A retrospective cohort analysis was conducted on 742 adult patients with lung cancer. In this retrospective study, all patients who underwent their first non-intubated VATS at Shandong Provincial Hospital Affiliated to Shandong First Medical University in Jinan, Shandong, China, between January 2024 to June 2025 were consecutively recruited. The inclusion criteria were: (I) age of 18–80 years old; (II) clinically diagnosed with lung cancer; (III) patients who underwent non-intubated VATS surgery; (IV) medical records with no significant missing data. The exclusion criteria were: (I) concurrent surgeries at other anatomical sites; (II) repeated hospitalizations within the study timeframe; (III) in-hospital mortality or self-discharge leading to treatment termination; (IV) disorders of the immune and/or hematological system; (V) conversion from thoracoscopy to open thoracotomy; (VI) participation in other clinical trials during the study period. Based on clinical experience, there are 26 potential factors influencing PLOS in non-intubated VATS patients with lung tumors. Assuming 20–25 samples per factor, the total required sample size was estimated at 520–650. Accounting for a 20% attrition rate, we calculated that 650–813 cases should be enrolled. Ultimately, this study enrolled 742 samples, fulfilling the research requirements. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of Shandong Provincial Hospital Affiliated to Shandong First Medical University (No. 2025-741). Individual consent for this retrospective analysis was waived.
Definition of primary indicators
LOS: duration from admission to hospital discharge, encompassing preoperative examination, postoperative care and observation.
PLOS: a patient’s hospitalization duration that significantly exceeds an expected benchmark for their clinical condition, diagnosis-related group (DRG), or standardized clinical pathway. PLOS is typically operationalized using statistical thresholds (e.g., 75th percentile, mean + 1 standard deviation) or clinically validated norms. In this study PLOS is defined as longer than the 75th percentile of the study population, while normal LOS was defined as LOS shorter than the 75th percentile of the study population (11).
-
Postoperative complications: postoperative complications were defined as specific adverse clinical events that newly occurred or worsened during the hospitalization for the index nonintubated VATS procedure and exerted a negative impact on patient recovery. Complications were classified according to severity using the Clavien-Dindo classification system, which categorizes postoperative adverse events into five grades (12): grade I (any deviation from the normal postoperative course without the need for pharmacological, surgical, endoscopic, or radiological interventions), grade II (requiring pharmacological treatment), grade III (requiring surgical, endoscopic, or radiological intervention), grade IV (life‑threatening complication requiring intensive‑care‑unit management), and grade V (death). Based on this system, complications were further grouped as:
Minor complications: Clavien-Dindo grades I–II (e.g., postoperative fever not requiring intervention, minor arrhythmia, pain controlled with analgesics).
Major complications: Clavien-Dindo grades III–V (e.g., hemorrhage necessitating re‑operation, prolonged air leak requiring chest‑tube drainage or intervention, pneumonia, respiratory failure).
The complications recorded and analyzed in this study included, but were not limited to: prolonged air leak (>5 days), arrhythmia, pulmonary infection, pleural effusion requiring drainage, hemorrhage, and pulmonary embolism. All complications were diagnosed based on integrated assessment of imaging studies, laboratory findings, and clinical documentation.
Statistical analysis
In this study, data were systematically collected through the Hospital Information System (HIS) and telephonic follow-ups. All statistical analyses were performed using SPSS version 26.0 and RStudio (R version 4.5.1). Descriptive statistics were used to summarize the study population. Continuous variables with a normal distribution were reported as mean ± standard deviation (SD) (x±s) and compared using independent-samples t-tests. Categorical variables were expressed as frequencies and percentages (%), with between-group comparisons conducted using the χ2 tests. For multi-level categorical variables [e.g., work status, body mass index (BMI) categories], an initial omnibus χ2 test was performed to assess overall group differences. If statistical significance was detected (P<0.05), Bonferroni-adjusted post-hoc pairwise comparisons (χ2 or Fisher’s exact tests) were conducted to identify specific inter-group disparities.
Variables that were statistically significant (P<0.05) in the univariate analyses were all entered simultaneously into a multivariable logistic regression model to identify independent predictors. From this model, coefficients were derived for each predictor. To construct a nomogram, the coefficients were scaled to assign points to each predictor. The largest coefficient was scaled to 100 points. For any predictor, the points assigned were calculated as: points = (regression coefficient * scaling factor). Where the Scaling Factor was chosen such that the largest coefficient corresponded to 100 points. For example, if a variable had a coefficient of 0.5 and the largest coefficient in the model was 1.0, that variable would be assigned 50 points (0.5/1.0 * 100). The total points for a patient were the sum of the points for each predictor. This total point score was then transformed into the probability of recurrence using the logistic regression equation. The reference category for categorical variables was assigned 0 points. The area under the receiver operating characteristic (ROC) curve was used to assess the discriminative ability of the model, with α =0.05 as the level of significance for testing.
Results
Patient characteristics
Consistent with previous literature (13), PLOS was defined as a hospital stay exceeding the 75th percentile of the cohort’s postoperative hospitalization duration. In the present study, the 75th percentile threshold was calculated to be 8 days. A total of 742 patients were included in the study and categorized into two groups: the PLOS group (LOS ≥8 days, n=216) and the normal length of stay (NLOS) group (LOS <8 days, n=526). The demographic, clinical, and perioperative characteristics of both groups are presented in Table 1.
Table 1. Clinical characteristics and univariate analysis of LOS.
| Variable | LOS ≥8 days (n=216) | LOS <8 days (n=526) | χ2/t | P |
|---|---|---|---|---|
| Sex | 7.284 | 0.007* | ||
| Male | 98 (45.4) | 183 (34.8) | ||
| Female | 118 (54.6) | 343 (65.2) | ||
| Marital status | 0.716 | 0.39 | ||
| Unmarried/divorced/bereaved | 13 (6.02) | 41 (7.79) | ||
| Married | 203 (93.98) | 485 (92.21) | ||
| Medical expense payment method | 6.031 | 0.049* | ||
| Self-pay | 1 (0.46) | 2 (0.38) | ||
| Resident medical insurance | 114 (52.78) | 226 (42.97) | ||
| Employee medical insurance | 101 (46.76) | 298 (56.65) | ||
| Long-term residence | 4.550 | 0.03* | ||
| Urban | 119 (55.09) | 334 (63.50) | ||
| Rural | 97 (44.91) | 192 (36.50) | ||
| Work status | 25.455 | <0.001* | ||
| Employed | 80 (37.04) | 302 (57.41) | ||
| Retired/unemployed | 136 (62.96) | 224 (42.59) | ||
| Number of diagnoses at discharge | 24.695 | <0.001* | ||
| 0–4 | 41 (18.98) | 192 (36.50) | ||
| 5–9 | 108 (50.00) | 230 (43.73) | ||
| ≥10 | 67 (31.02) | 104 (19.77) | ||
| Smoking history | 14.103 | <0.001* | ||
| Yes | 71 (32.87) | 105 (19.96) | ||
| No | 145 (67.13) | 421 (80.04) | ||
| Alcohol consumption history | 5.127 | 0.02* | ||
| Yes | 53 (24.54) | 91 (17.30) | ||
| No | 163 (75.46) | 435 (82.70) | ||
| Preoperative chronic conditions | 2.820 | 0.09 | ||
| No | 49 (22.69) | 151 (28.71) | ||
| Yes | 167 (77.31) | 375 (71.29) | ||
| BMI (kg/m2) | 10.770 | 0.01* | ||
| <18.5 | 9 (4.17) | 7 (1.33) | ||
| 18.5–24.9 | 131 (60.65) | 284 (53.99) | ||
| 25.0–29.9 | 68 (31.48) | 202 (38.40) | ||
| ≥30.0 | 8 (3.70) | 33 (6.27) | ||
| Reason for visit | 6.982 | 0.03* | ||
| Physical examination | 194 (89.81) | 500 (95.06) | ||
| Pulmonary symptoms | 14 (6.48) | 16 (3.04) | ||
| Other | 8 (3.70) | 10 (1.90) | ||
| Neoadjuvant therapy | 0.6471 | 0.42 | ||
| Yes | 3 (1.39) | 4 (0.76) | ||
| No | 213 (98.61) | 522 (99.24) | ||
| Tumor pathology | 26.019 | <0.001* | ||
| Carcinoma in situ | 24 (11.11) | 124 (23.57) | ||
| Microinvasive carcinoma | 50 (23.15) | 158 (30.04) | ||
| Invasive carcinoma | 114 (52.78) | 189 (35.93) | ||
| Others | 28 (12.96) | 55 (10.46) | ||
| Tumor histological classification | 6.52 | 0.08 | ||
| Adenocarcinoma | 174 (80.56) | 453 (86.12) | ||
| Squamous cell carcinoma | 10 (4.63) | 9 (1.71) | ||
| Mucinous carcinoma | 6 (2.78) | 13 (2.47) | ||
| Others | 26 (12.04) | 51 (9.70) | ||
| T component of the TNM staging | 21.545 | <0.001* | ||
| T0 | 18 (8.33) | 44 (8.37) | ||
| Tis | 25 (11.57) | 126 (23.95) | ||
| T1 | 154 (71.30) | 338 (64.26) | ||
| T2 and above | 19 (8.80) | 18 (3.42) | ||
| N component of the TNM staging | 8.870 | 0.01* | ||
| N0 | 201 (93.06) | 513 (97.53) | ||
| N1 | 10 (4.63) | 10 (1.90) | ||
| N2 | 5 (2.31) | 3 (0.57) | ||
| Tumor location | 4.788 | 0.054 | ||
| Central | 9 (4.17) | 8 (1.52) | ||
| Peripheral | 207 (95.83) | 518 (98.48) | ||
| Surgical methods | 25.167 | <0.001* | ||
| Lobectomy | 51 (23.61) | 54 (10.27) | ||
| Wedge resection | 111 (51.39) | 324 (61.60) | ||
| Segmentectomy | 32 (14.81) | 105 (19.96) | ||
| Combined resection | 14 (6.48) | 30 (5.70) | ||
| Others | 8 (3.70) | 13 (2.47) | ||
| Chest entry approach | 3.277 | 0.35 | ||
| 3rd intercostal space, mid-axillary line | 9 (4.17) | 34 (6.46) | ||
| 4th intercostal space, mid-axillary line | 123 (56.94) | 296 (56.27) | ||
| 5th intercostal space, mid-axillary line | 82 (37.96) | 184 (34.98) | ||
| Others | 2 (0.93) | 12 (2.28) | ||
| CT-guided localization | 0.662 | 0.41 | ||
| Yes | 43 (19.91) | 119 (22.62) | ||
| No | 173 (80.09) | 407 (77.38) | ||
| Chest tube drainage volume (mL) | 100.631 | <0.001* | ||
| ≤200 | 106 (49.07) | 431 (81.94) | ||
| 201–400 | 58 (26.85) | 77 (14.64) | ||
| >400 | 52 (24.07) | 18 (3.42) | ||
| Chest tube indwelling time (h) | 130.903 | <0.001* | ||
| 0–24 | 16 (7.41) | 131 (24.90) | ||
| 25–48 | 113 (52.31) | 358 (68.06) | ||
| >48 | 87 (40.28) | 37 (7.03) | ||
| ASA physical status score | 15.533 | <0.001* | ||
| 1 | 0 (0.00) | 3 (0.57) | ||
| 2 | 189 (87.50) | 498 (94.68) | ||
| 3 | 25 (11.57) | 24 (4.56) | ||
| 4 | 2 (0.93) | 1 (0.19) | ||
| Complications | 61.935 | <0.001* | ||
| Yes | 105 (48.61) | 105 (19.96) | ||
| No | 111 (51.39) | 421 (80.04) | ||
| Postoperative pain score | 0.0576 | 0.81 | ||
| 1–3 | 210 (97.22) | 513 (97.53) | ||
| 4–6 | 6 (2.78) | 13 (2.47) | ||
| Reintubation | 22.186 | <0.001* | ||
| Yes | 9 (4.17) | 0 (0.00) | ||
| No | 207 (95.83) | 526 (100.00) | ||
| Age (years) | 61.76±9.37 | 55.36±11.12 | 7.433 | <0.001* |
| Hospitalization cost (CNY) | 40,317.51±6,801.60 | 35,565.41±4,927.25 | 10.619 | <0.001* |
| Preoperative wait time (days) | 4.72±1.82 | 2.83±0.87 | 19.096 | <0.001* |
| Operative time (minutes) | 67.00±30.15 | 54.78±23.44 | 5.913 | <0.001* |
| Surgical blood loss (mL) | 20.01±15.56 | 15.88±10.13 | 4.272 | <0.001* |
| Number of lymph nodes dissected | 4.66±3.79 | 3.50±2.96 | 4.473 | <0.001* |
| Number of lymph node stations dissected | 3.49±1.68 | 3.08±2.05 | 2.555 | <0.01* |
| Anesthesia duration (min) | 98.49±32.99 | 84.19±25.53 | 6.341 | <0.001* |
| Anesthesia recovery time (min) | 51.13±17.36 | 49.77±13.56 | 1.147 | 0.25 |
Data are presented as mean ± standard deviation or n (%). *, statistical significance (P<0.05). ASA, American Society of Anesthesiologists; BMI, body mass index; CT, computed tomography; LOS, length of stay; TNM, tumor node metastasis.
Risk of PLOS
Univariate analysis of PLOS in lung cancer patients
Compared with the normal LOS group, patients in the PLOS group were more likely to be male (P=0.007), retired or unemployed (P<0.001), and reside in rural areas (P=0.03). They had a greater number of comorbidities (P<0.001), a higher prevalence of smoking (P<0.001) and a history of alcohol consumption (P=0.02), as well as higher American Society of Anesthesiologists (ASA) scores (P<0.001). Pathologically, the PLOS group had a significantly higher proportion of invasive carcinoma (P<0.001). Surgically, these patients more frequently underwent lobectomy (P<0.001). Postoperatively, the PLOS group demonstrated significantly higher chest tube drainage volumes (P<0.001), longer chest tube durations (P<0.001), a higher complication rate (48.6% vs. 20.0%, P<0.001), and a need for reintubation in some cases (4.2% vs. 0%, P<0.001). Additionally, patients with PLOS were significantly older, had longer times from admission to surgery, longer operative and anesthesia durations, greater intraoperative blood loss, dissection of more lymph nodes, and incurred higher hospitalization costs (all P<0.001). No significant differences between groups were observed for marital status, neoadjuvant therapy, tumor location, or postoperative pain scores, see Table 1 for details.
In addition, for categorical variables with three or more categories, multiple comparisons were performed in order to identify true predictors. Significant differences were found for BMI <18.5 kg/m2 compared to 18.5–24.9 kg/m2 (χ2=4.280, P=0.03), 25.0–29.9 kg/m2 (χ2=7.409, P=0.006), and ≥30.0 kg/m2 (χ2=7.422, P=0.006). For chest tube drainage volume, all comparisons showed significant differences (≤200 vs. 201–400 mL: χ2=31.538, P<0.001; ≤200 vs. >400 mL: χ2=95.695, P<0.001; 201–400 vs. >400 mL: χ2=18.188, P<0.001). Other details are shown in Table 2.
Table 2. Results of post-hoc pairwise comparisons for univariate analysis of LOS.
| No. | Variable | LOS ≥8 days (n=216), n | LOS <8 days (n=526), n | χ2 | P |
|---|---|---|---|---|---|
| Medical expense payment method | |||||
| ① | Self-pay | 1 | 2 | 0.000 | 0.99 |
| Resident medical insurance | 114 | 226 | |||
| ② | Self-pay | 1 | 2 | 0.101 | 0.75 |
| Employee medical insurance | 101 | 298 | |||
| ③ | Resident medical insurance | 114 | 226 | 6.007 | 0.01* |
| Employee medical insurance | 101 | 298 | |||
| Number of diagnoses at discharge | |||||
| ① | 0–4 | 41 | 192 | 14.739 | <0.001* |
| 5–9 | 108 | 230 | |||
| ② | 0–4 | 41 | 192 | 23.459 | <0.001* |
| ≥10 | 67 | 104 | |||
| ③ | 5–9 | 108 | 230 | 2.630 | 0.10 |
| ≥10 | 67 | 104 | |||
| BMI (kg/m2) | |||||
| ① | <18.5 | 9 | 7 | 4.280 | 0.03* |
| 18.5–24.9 | 131 | 284 | |||
| ② | <18.5 | 9 | 7 | 7.409 | 0.006* |
| 25.0–29.9 | 68 | 202 | |||
| ③ | <18.5 | 9 | 7 | 7.422 | 0.006* |
| ≥30.0 | 8 | 33 | |||
| ④ | 18.5–24.9 | 131 | 284 | 3.231 | 0.07 |
| 25.0–29.9 | 68 | 202 | |||
| ⑤ | 18.5–24.9 | 131 | 284 | 2.559 | 0.11 |
| ≥30.0 | 8 | 33 | |||
| ⑥ | 25.0–29.9 | 68 | 202 | 0.620 | 0.43 |
| ≥30.0 | 8 | 33 | |||
| Reason for visit | |||||
| ① | Physical examination | 194 | 500 | 4.918 | 0.02* |
| Pulmonary symptoms | 14 | 16 | |||
| ② | Physical examination | 194 | 500 | 2.348 | 0.12 |
| Others | 8 | 10 | |||
| ③ | Pulmonary symptoms | 14 | 16 | 0.022 | 0.88 |
| Others | 8 | 10 | |||
| Tumor pathology | |||||
| ① | Carcinoma in situ | 24 | 124 | 3.213 | 0.07 |
| Microinvasive carcinoma | 50 | 158 | |||
| ② | Carcinoma in situ | 24 | 124 | 21.458 | <0.001* |
| Invasive carcinoma | 114 | 189 | |||
| ③ | Carcinoma in situ | 24 | 124 | 9.356 | 0.002* |
| Others | 28 | 55 | |||
| ④ | Microinvasive carcinoma | 50 | 158 | 10.445 | 0.001* |
| Invasive carcinoma | 114 | 189 | |||
| ⑤ | Microinvasive carcinoma | 50 | 158 | 2.843 | 0.09 |
| Others | 28 | 55 | |||
| ⑥ | Invasive carcinoma | 114 | 189 | 0.424 | 0.51 |
| Others | 28 | 55 | |||
| T component of the TNM staging | |||||
| ① | T0 | 18 | 44 | 4.246 | 0.03* |
| Tis | 25 | 126 | |||
| ② | T0 | 18 | 44 | 0.132 | 0.71 |
| T1 | 154 | 338 | |||
| ③ | T0 | 18 | 44 | 4.932 | 0.02* |
| T2 and above | 19 | 18 | |||
| ④ | Tis | 25 | 126 | 12.504 | <0.001* |
| T1 | 154 | 338 | |||
| ⑤ | Tis | 25 | 126 | 20.070 | <0.001* |
| T2 and above | 19 | 18 | |||
| ⑥ | T1 | 154 | 338 | 6.286 | 0.01* |
| T2 and above | 19 | 18 | |||
| N component of the TNM staging | |||||
| ① | N0 | 201 | 513 | 4.534 | 0.03* |
| N1 | 10 | 10 | |||
| ② | N0 | 201 | 513 | 4.578 | 0.03* |
| N2 | 5 | 3 | |||
| ③ | N1 | 10 | 10 | 0.359 | 0.54 |
| N2 | 5 | 3 | |||
| Surgical methods | |||||
| ① | Lobectomy | 51 | 54 | 21.407 | <0.001* |
| Wedge resection | 111 | 324 | |||
| ② | Lobectomy | 51 | 54 | 16.770 | <0.001* |
| Segmentectomy | 32 | 105 | |||
| ③ | Lobectomy | 51 | 54 | 3.539 | 0.06 |
| Combined resection | 14 | 30 | |||
| ④ | Lobectomy | 51 | 54 | 0.771 | 0.38 |
| Others | 8 | 13 | |||
| ⑤ | Wedge resection | 111 | 324 | 0.259 | 0.61 |
| Segmentectomy | 32 | 105 | |||
| ⑥ | Wedge resection | 111 | 324 | 0.823 | 0.36 |
| Combined resection | 14 | 30 | |||
| ⑦ | Wedge resection | 111 | 324 | 1.643 | 0.20 |
| Others | 8 | 13 | |||
| ⑧ | Segmentectomy | 32 | 105 | 1.258 | 0.26 |
| Combined resection | 14 | 30 | |||
| ⑨ | Segmentectomy | 32 | 105 | 2.092 | 0.14 |
| Others | 8 | 13 | |||
| ⑩ | Combined resection | 14 | 30 | 0.250 | 0.61 |
| Others | 8 | 13 | |||
| Chest tube drainage volume (mL) | |||||
| ① | ≤200 | 106 | 431 | 31.538 | <0.001* |
| 201–400 | 58 | 77 | |||
| ② | ≤200 | 106 | 431 | 95.695 | <0.001* |
| >400 | 52 | 18 | |||
| ③ | 201–400 | 58 | 77 | 18.188 | <0.001* |
| >400 | 52 | 18 | |||
| Chest tube indwelling time (h) | |||||
| ① | 0–24 | 16 | 131 | 11.653 | <0.001* |
| 25–48 | 113 | 358 | |||
| ② | 0–24 | 16 | 131 | 100.307 | <0.001* |
| >48 | 87 | 37 | |||
| ③ | 25–48 | 113 | 358 | 93.767 | <0.001* |
| >48 | 87 | 37 | |||
| ASA physical status score | |||||
| ① | 1 | 0 | 3 | 1.137 | 0.28 |
| 2 | 189 | 498 | |||
| ② | 1 | 0 | 3 | 2.948 | 0.08 |
| 3 | 25 | 24 | |||
| ③ | 1 | 0 | 3 | 3.000 | 0.08 |
| 4 | 2 | 1 | |||
| ④ | 2 | 189 | 498 | 12.258 | <0.001* |
| 3 | 25 | 24 | |||
| ⑤ | 2 | 189 | 498 | 2.288 | 0.13 |
| 4 | 2 | 1 | |||
| ⑥ | 3 | 25 | 24 | 0.277 | 0.59 |
| 4 | 2 | 1 | |||
*, statistical significance (P<0.05). ASA, American Society of Anesthesiologists; BMI, body mass index; LOS, length of stay; TNM, tumor node metastasis.
Logistic regression-based screening variables and model construction
Variables with statistically significant coefficients identified through univariate regression analysis were subsequently incorporated into the multivariable logistic regression model to identify several independent predictors of PLOS (LOS ≥8 days), The assignment of values for the independent variables was as follows: BMI: <18.5 kg/m2 =1, 18.5–24.9 kg/m2 =2, 25.0–29.9 kg/m2 =3; 30.0 kg/m2 and above =4; chest tube drainage volume: 200 mL and below =1, 201–400 mL =2, more than 400 mL =3; chest tube indwelling time: 0–24 h =1, 25–48 h =2, more than 48 h =3; complications: no complications =1, have complications =2. The results showed that age, preoperative wait time, BMI (30.0 kg/m2 and above), chest tube drainage volume, chest tube indwelling time (more than 48 h) and complications were factors influencing the PLOS in patients undergoing non-intubated thoracoscopic resection of lung cancer in patients (P<0.05), as shown in Table 3.
Table 3. Multivariable logistic regression analysis of length of stay.
| Predictor | Coefficient | Standard error | Wald χ2 | P | OR | 95% confidence interval of OR |
|---|---|---|---|---|---|---|
| Intercept | 9.689 | 1.343 | 52.071 | <0.001 | 0.000 | 0.000, 0.000 |
| Age | 0.052 | 0.015 | 11.696 | <0.001 | 1.053 | 1.022, 1.085 |
| Preoperative wait time | 2.045 | 0.184 | 124.055 | <0.001 | 7.729 | 5.393, 11.076 |
| BMI (kg/m2) | ||||||
| 18.5–24.9† | −1.953 | 1.010 | 3.744 | 0.053 | 0.142 | 0.020, 1.026 |
| 25.0–29.9† | −1.979 | 1.014 | 3.810 | 0.051 | 0.138 | 0.019, 1.008 |
| 30.0 and above† | −3.137 | 1.209 | 6.739 | 0.009 | 0.043 | 0.004, 0.464 |
| Chest tube drainage volume (mL) | ||||||
| 201–400‡ | 0.970 | 0.342 | 8.049 | 0.005 | 2.637 | 1.349, 5.154 |
| More than 400 | 2.422 | 0.549 | 19.466 | <0.001 | 11.266 | 3.842, 33.038 |
| Chest tube indwelling time (h) | ||||||
| 25–48§ | 0.169 | 0.406 | 0.172 | 0.67 | 1.184 | 0.534, 2.625 |
| More than 48§ | 2.048 | 0.546 | 14.055 | <0.001 | 7.755 | 2.658, 22.629 |
| Complications | ||||||
| Yes¶ | 1.089 | 0.310 | 12.362 | <0.001 | 2.970 | 1.619, 5.449 |
Coefficients represent the log-odds of “length of stay = yes” compared to “length of stay = no”. †, BMI <18.5 kg/m2; ‡, 200 mL and below; §, 0–24 h; ¶, complications = no. BMI, body mass index; OR, odds ratio.
Older age [odds ratio (OR) =1.053, P<0.001] and longer preoperative wait time (OR =7.729, P<0.001) were significantly associated with an increased risk of PLOS. Compared with underweight patients (BMI <18.5 kg/m2), those with BMI ≥30.0 kg/m2 had a 23-fold higher odds of PLOS (OR =0.043, P=0.009) , whereas overweight categories (BMI 18.5–24.9 and 25.0–29.9 kg/m2) showed non-significant trends toward increased risk (P=0.053 and 0.051, respectively). Chest tube drainage volumes exceeding 200 mL were associated with markedly increased PLOS odds (201–400 mL: OR =2.637; >400 mL: OR =11.266; both P≤0.005), as was chest tube indwelling time >48 h (OR =7.755, P<0.001). Clearly, patients with complications (OR =2.970, P<0.001) exhibited a higher PLOS risk compared to those without complications.
Based on these findings, a predictive nomogram was developed using the R language, which visually illustrates the influence of various risk factors on the risk of PLOS in patients undergoing non-intubated thoracoscopic resection of lung cancer in patients (Figure 1). The area under the curve (AUC) for this model was 0.943 (Figure 2), demonstrating excellent discriminative ability. The 95% confidence interval (CI) ranged from 0.926 to 0.961, with a maximum Youden index of 0.748. The optimal cutoff value was determined to be 0.389, which further validated the stability of the model’s predictive performance. With a sensitivity of 0.824 and a specificity of 0.924, the model exhibited robust performance in accurately identifying both true positives and true negatives.
Figure 1.
Nomogram for prolonged length of stay. Referent: “BMI <18.5 kg/m2”, chest tube drainage volume: “200 mL and below”, chest tube indwelling time: “0–24 h”, Complications: “no complications”. BMI, body mass index.
Figure 2.
Receiver operating characteristic curve for the prediction model. The area under the curve is 0.943.
Discussion
Lung cancer represents a major public health challenge in China, characterized by a high incidence rate and a substantial disease burden. Given the constraints on healthcare resources, rationally accelerating bed turnover and reducing the LOS—while maintaining the quality of care—not only alleviates the financial burden on patients but also improves the efficiency of healthcare resource allocation. Such measures are essential to ensuring timely diagnosis and treatment for a greater number of patients.
Lung cancer studies have demonstrated significant variations in postoperative LOS across different healthcare systems. In an analysis of 771 patients undergoing lobectomy, Li et al. reported a median LOS of 9.0 days in the PLOS group, which exceeds the finding of the present study (14). Comparative other studies indicate median hospitalization durations of 3–8 and 5 days, respectively, highlighting substantial international differences (15,16). Furthermore, multiple study has shown that non-intubated anesthesia significantly reduces postoperative hospitalization duration, operative time, postoperative recovery time, thoracic drainage tube duration, and chest tube drainage volume compared with conventional endotracheal intubation anesthesia (17).
In this cohort, the mean hospitalization duration was 8 days. Future studies incorporating a comparison group are needed to determine whether predictive model can effectively reduce LOS. PLOS may be attributed to perioperative management protocols and complication profiles, warranting further investigation to identify relevant determinants. Systematic analysis of perioperative clinical data to identify PLOS-related predictors could optimize clinical management and facilitate early intervention strategies. Such an approach would effectively mitigate risks associated with either excessively long hospitalizations or premature discharge.
The duration of hospitalization serves as a critical indicator of patient recovery. PLOS typically reflects delayed recovery and is associated with an elevated risk of complications and potentially poorer long-term outcomes. Wright et al. identified age, gender, Zubrod performance status, and comorbidities as significant contributors to PLOS (17). Corroborating these findings, our study establishes that age, postoperative complications, drainage duration, preoperative waiting time, and drainage volume also serve as independent predictors of PLOS.
Age emerges as a pivotal determinant of PLOS. Both the current study and the investigation by Hu et al. revealed progressively extended hospitalization durations in patients over 50 years old, with particularly pronounced increases observed in those exceeding 70 years of age (18). Multiple studies further corroborate that age >70 years independently predicts PLOS in elderly lung cancer patients (19-21). These collective finding underscore the necessity for enhanced clinical vigilance in elderly populations undergoing non-intubated thoracoscopic lung surgery. Advanced age is associated with reduced physiological reserve, compromised immune function, and increased susceptibility to postoperative complications, all of which collectively impede recovery trajectories.
Notably, elderly patients frequently present with comorbidities such as malnutrition and sarcopenia. These conditions compromise surgical tolerance and impair tissue repair, thereby delaying postoperative recovery and prolonging the LOS. Evidence from previous study confirms a well-established association between low BMI and an increased risk of malnutrition, which in turn exacerbates postoperative mortality and complication rates (18). Consequently, the implementation of comprehensive preoperative assessment protocols for elderly patients is imperative. Such protocols should encompass the management of comorbidities, nutritional optimization, and functional status evaluation. These multidisciplinary strategies enable individualized treatment planning, facilitate accelerated recovery pathways, and ultimately reduce hospitalization duration.
This study found that 48.61% of patients with PLOS experienced postoperative complications. This observation reinforces the established clinical understanding that postoperative complications serve as a primary driver of extended hospitalization. Consequently, complication prevention should be a central strategy for reducing PLOS. To achieve this, future research should prioritize identifying upstream, modifiable risk factors—such as preoperative frailty or variations in surgical technique—that predispose patients to complications. Targeting these factors represents the most effective intervention point for shortening hospital stay by addressing the root causes of prolonged recovery. Our finding align with the existing literature. For instance, study by Farjah et al. have consistently demonstrated a significant association between postoperative complications and PLOS (22). Farjah et al. notably reported a complication rate as high as 99% among patients with extended hospitalization, further substantiating the role of complications as a key determinant of LOS (22). Furthermore, a retrospective cohort study corroborated a positive correlation between postoperative complications and PLOS following lung cancer surgery (23). In our cohort, respiratory complications were predominant, including pneumonia, acute respiratory distress syndrome (ARDS), atelectasis, and prolonged air leak. Other frequent events involved cardiovascular and cerebrovascular systems, such as arrhythmia, myocardial infarction, cerebral infarction, and deep vein thrombosis. Of particular note, prolonged air leak was one of the most frequently observed complications.
Evidence indicates that a prolonged air leak is a well-documented driver of extended hospitalization, increased healthcare costs, and a higher risk of cardiopulmonary complications, all of which adversely impact patient prognosis (24). Study have shown that preoperative prehabilitation for lung cancer patients can reduce postoperative hospital stay, decrease postoperative pulmonary complications, and lower the incidence of pulmonary air leaks (25). Therefore, preoperative interventions—including smoking cessation guidance, nutritional support, and respiratory prehabilitation—should be prioritized during the critical preoperative window to mitigate this complication. Multidisciplinary collaboration is essential for early risk stratification and the implementation of targeted preventive strategies, which are key to optimizing clinical outcomes and alleviating the burdens associated with PLOS.
The preoperative waiting time, defined as the interval from hospital admission to surgery, typically spans 1–2 days to allow for essential evaluations and preparations, with most operations scheduled between days 2 and 4 (26). This study identified a prolonged preoperative waiting time as an independent risk factor for PLOS. The reasons for such delays can be multifactorial, including the need for further optimization of blood pressure or glycemic control, a poor physiological status requiring multidisciplinary consultation, delayed completion of preoperative tests, or institutional scheduling conflicts, such as those caused by holidays. Extended waiting periods may contribute to physical deconditioning, increase surgical risk, and hinder postoperative recovery, potentially due to environmental stress and preoperative anxiety (27). To facilitate recovery and shorten LOS, healthcare institutions should focus on minimizing non-essential delays and improving scheduling efficiency. For patients facing unavoidably long waits, intensified monitoring and proactive measures to prevent complications are crucial. Notably, the implementation of “pre-admission” protocols, where preoperative assessments are completed during outpatient visits, has proven effective in many hospitals. This approach can reduce financial burdens, shorten hospital stays, alleviate patient anxiety, and improve the utilization of healthcare resources.
Although operative duration was not retained in the final regression model of this study, its influence on the LOS in patients undergoing non-intubated thoracoscopic surgery for lung cancer remains to be fully elucidated. However, extensive evidence from other surgical domains, including colorectal cancer, endometrial cancer, joint replacement, and spinal deformity procedures, consistently demonstrates a significant association between prolonged operative time and extended LOS (28,29). Furthermore, extended surgical duration is an established risk factor for several complications, such as surgical site infections, pneumonia, atelectasis, reintubation, and unplanned intensive care unit (ICU) admission, all of which can indirectly prolong hospitalization (30,31).
The prediction model has practical value
The predictive model developed in this study demonstrates substantial clinical utility, exhibiting satisfactory calibration and discrimination (AUC: 0.943). The calibration curve showed good agreement between predicted probabilities and observed outcomes, thus supporting its validity. We further constructed a nomogram to visualize the individual contribution of each risk factor to the probability of PLOS. Due to its parsimonious structure and strong clinical applicability, this tool facilitates the identification of key variables driving prolonged hospitalization. These findings suggest that precise risk stratification using this model, followed by targeted perioperative interventions, could optimize patient management, mitigate the risk of extended stays, and improve therapeutic efficiency and overall outcomes in lung cancer surgery. Consequently, this model holds promise as a routine screening tool in thoracic surgery departments, enabling early identification of high-risk patients, rational allocation of medical resources, and a potential reduction in healthcare costs.
Limitations
There are several limitations in this study. First, its single-center retrospective design may introduce selection bias and limit the generalizability of the finding, necessitating future external validation. Second, although significant clinical variables were included, some potential confounders, such as detailed surgical complexity and socio-economic factors, were not considered. While this study identified significant associations, it cannot definitively establish causality. Further prospective studies are needed to confirm these relationships. Third, the duration of our study was short, with the 18-month period potentially restricting sample diversity, and seasonal biases might have been introduced. Fourth, In the present study, we defined PLOS as the 75th percentile of the cohort’s hospital stay duration (8 days). Although this is a common data-driven approach in the development of predictive model, it limits the direct comparability of our finding with study that employ alternative definitions, such as those based on clinical pathway benchmarks, DRG expected lengths of stay, or international standards. Future research should aim to define PLOS based on clinical outcomes (e.g., timing of complication onset) or consensus-driven, multi-center criteria, thereby enhancing the generalizability and clinical utility of predictive model.
Conclusions
In conclusion, this study developed a clinically applicable prediction model for PLOS following non-intubated VATS lung cancer surgery. The model, incorporating key predictors like age and complications, serves as a practical tool for early risk stratification. It holds significant promise for guiding personalized care and optimizing clinical pathways to enhance recovery and resource utilization.
Supplementary
The article’s supplementary files as
Acknowledgments
None.
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. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of Shandong Provincial Hospital Affiliated to Shandong First Medical University (No. 2025-741). Individual consent for this retrospective analysis was waived.
Footnotes
Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2240/rc
Funding: None.
Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2240/coif). The authors have no conflicts of interest to declare.
Data Sharing Statement
Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2240/dss
References
- 1.Yang QH, Hu YX, Ni PW, et al. Trends of incidence and mortality of lung cancer in China from 1990 to 2021 and the prediction for 2035. Chinese General Practice 2026;29:256-63. [Google Scholar]
- 2.Tong Y, Wu J, Wu X, et al. Analgesic Efficacy of Thoracoscopic Direct-View Versus Ultrasound-Guided Thoracic Paravertebral Block in Multi-Port Video-Assisted Thoracoscopic Lung Surgery: A Randomized Controlled Non-Inferiority Study. Drug Des Devel Ther 2025;19:1825-38. 10.2147/DDDT.S492040 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Dittberner FA, Bendixen M, Licht PB. Long-term survival from a randomized controlled trial of lobectomy by video-assisted thoracoscopic surgery versus thoracotomy for early-stage lung cancer. Eur J Cardiothorac Surg 2025;67:ezaf017. 10.1093/ejcts/ezaf017 [DOI] [PubMed] [Google Scholar]
- 4.Luo T, Zhang Y, Xu BW, et al. Non-intubated vs. intubated video-assisted thoracoscopic surgery for the treatment of thoracic diseases: a systematic review and meta-analysis of propensity score-matched cohorts. Front Surg 2025;12:1661466. 10.3389/fsurg.2025.1661466 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Elmaleh YD, Guessous K, Zanoun N. Advancing thoracic surgery: the clinical, economic, and environmental promise of nonintubated video-assisted thoracoscopic surgery (NIVATS). Br J Anaesth 2025;135:1566-8. 10.1016/j.bja.2025.07.055 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Guo Y, Wang F, Ma S, et al. Relationship between atherogenic index of plasma and length of stay in critically ill patients with atherosclerotic cardiovascular disease: a retrospective cohort study and predictive modeling based on machine learning. Cardiovasc Diabetol 2025;24:95. 10.1186/s12933-025-02654-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zhang Z, Li J, Wang W, et al. Predictors and outcomes associated with prolonged hospital length of stay in intracerebral hemorrhage: a multicenter prospective cohort study in China. BMC Med 2025;23:404. 10.1186/s12916-025-04241-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Pezeshkian F, Leo R, McAllister MA, et al. Predictors of prolonged hospital stay after segmentectomy. J Thorac Cardiovasc Surg 2025;169:420-6. 10.1016/j.jtcvs.2024.04.021 [DOI] [PubMed] [Google Scholar]
- 9.Kendall F, Silva G, Drummond M, et al. Predictors of prolonged hospital stay in patients undergoing lung resection. Disabil Rehabil 2024;46:5220-6. 10.1080/09638288.2023.2297936 [DOI] [PubMed] [Google Scholar]
- 10.Kong W, Wang H, Chen Y, et al. Association between length of stay and postoperative survival in patients with lung cancer: a propensity score matching analysis based on National Cancer Database. J Thorac Dis 2023;15:1785-93. 10.21037/jtd-23-407 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hurisa Dadi H, Habte N, Mulu Y. Length of hospital stay and associated factors among adult surgical patients admitted to surgical wards in Amhara Regional State Comprehensive Specialized Hospitals, Ethiopia. PLoS One 2024;19:e0296143. 10.1371/journal.pone.0296143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zeng K, Wang QZ, Li YL. Application of Clavien-Dindo classification system to complications of minimally-invasive percutaneous nephrolithotomy. China Journal of Modern Medicine 2016;26:96-100. [Google Scholar]
- 13.Grigor EJM, Ivanovic J, Anstee C, et al. Impact of Adverse Events and Length of Stay on Patient Experience After Lung Cancer Resection. Ann Thorac Surg 2017;104:382-8. 10.1016/j.athoracsur.2017.05.025 [DOI] [PubMed] [Google Scholar]
- 14.Li HW, Zhang LY. Analysis of risk factors for prolonged hospital stay in lung cancer patients after lobectomy. Chinese Journal of Clinical Thoracic and Cardiovascular Surgery 2019;26:425-30. [Google Scholar]
- 15.von Meyenfeldt EM, Hoeijmakers F, Marres GMH, et al. Variation in length of stay after minimally invasive lung resection: a reflection of perioperative care routines? Eur J Cardiothorac Surg 2020;57:747-53. 10.1093/ejcts/ezz303 [DOI] [PubMed] [Google Scholar]
- 16.Tou S, Matsumoto K, Hashinokuchi A, et al. Identifying Key Variances in Clinical Pathways Associated With Prolonged Hospital Stays Using Machine Learning and ePath Real-World Data: Model Development and Validation Study. JMIR Med Inform 2025;13:e71617. 10.2196/71617 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Wright CD, Gaissert HA, Grab JD, et al. Predictors of prolonged length of stay after lobectomy for lung cancer: a Society of Thoracic Surgeons General Thoracic Surgery Database risk-adjustment model. Ann Thorac Surg 2008;85:1857-65; discussion 1865. 10.1016/j.athoracsur.2008.03.024 [DOI] [PubMed] [Google Scholar]
- 18.Hu XL, Xu ST, Wang XC, et al. Development and validation of nomogram estimating post-surgery hospital stay of lung cancer patients: relevance for predictive, preventive, and personalized healthcare strategies. EPMA J 2019;10:173-83. 10.1007/s13167-019-00168-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Healy GL, Stuart CM, Dyas AR, et al. Association between postoperative complications and hospital length of stay: a large-scale observational study of 4,495,582 patients in the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) registry. Patient Saf Surg 2024;18:29. 10.1186/s13037-024-00409-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Jawitz OK, Wang Z, Boffa DJ, et al. The differential impact of preoperative comorbidity on perioperative outcomes following thoracoscopic and open lobectomies. Eur J Cardiothorac Surg 2017;51:169-74. 10.1093/ejcts/ezw239 [DOI] [PubMed] [Google Scholar]
- 21.Huang L, Kehlet H, Petersen RH. Why do patients stay in hospital after enhanced recovery thoracoscopic wedge resection? A prospective observational study. Eur J Cardiothorac Surg 2024;66:ezae366. 10.1093/ejcts/ezae366 [DOI] [PubMed] [Google Scholar]
- 22.Farjah F, Lou F, Rusch VW, et al. The quality metric prolonged length of stay misses clinically important adverse events. Ann Thorac Surg 2012;94:881-7; discussion 887-8. 10.1016/j.athoracsur.2012.04.082 [DOI] [PubMed] [Google Scholar]
- 23.Xu XY, Mo XX, Zhang S, et al. A review of prehabilitation before surgery in patients with lung cancer and chronic obstructive pulmonary disease. Journal of Nursing Science 2024;39:25-9. [Google Scholar]
- 24.Yu CC, Chen JR, Chen RY, et al. Effect of three-dimensional psychological intervention combined with postural intervention on mental health and rehabilitation in patients with pneumothorax after thoracoscopic lung cancer surgery. China Journal of Health Psychology 2025;33:86-91. [Google Scholar]
- 25.Zhang Z, Mostofian F, Ivanovic J, et al. All grades of severity of postoperative adverse events are associated with prolonged length of stay after lung cancer resection. J Thorac Cardiovasc Surg 2018;155:798-807. 10.1016/j.jtcvs.2017.09.094 [DOI] [PubMed] [Google Scholar]
- 26.Zhang D, Zhang T, Jiang S, et al. Influencing factors of preoperative hospital stay prolongation in elderly lung cancer patients. Chinese Journal of Multiple Organ Diseases in the Elderly 2024;23:212-6. [Google Scholar]
- 27.Le X, Zhao TY, Yu YH, et al. Anxiety status and influencing factors of surgical patients during preoperative waiting period. Chinese Nursing Management 2017;17:886-92. [Google Scholar]
- 28.Shen X, Zhou C, Hua Q, et al. Impact of operation duration on short-term and long-term prognosis in patients undergoing radical colorectal surgery. J Cancer 2022;13:1160-7. 10.7150/jca.65817 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Li H, Jiao J, Zhang S, et al. Construction and Comparison of Predictive Models for Length of Stay after Total Knee Arthroplasty: Regression Model and Machine Learning Analysis Based on 1,826 Cases in a Single Singapore Center. J Knee Surg 2022;35:7-14. 10.1055/s-0040-1710573 [DOI] [PubMed] [Google Scholar]
- 30.Wang Z, Pang L, Tang J, et al. Video-assisted thoracoscopic surgery versus muscle-sparing thoracotomy for non-small cell lung cancer: a systematic review and meta-analysis. BMC Surg 2019;19:144. 10.1186/s12893-019-0618-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Motono N, Mizoguchi T, Ishikawa M, et al. Analysis of risk factors of postoperative complication for non-small cell lung cancer. BMC Pulm Med 2024;24:333. 10.1186/s12890-024-03054-1 [DOI] [PMC free article] [PubMed] [Google Scholar]