Predictive accuracy of the Kristersson/Olsen formula after lung resection: a prospective study comparing minimally invasive surgery and thoracotomy

Abstract

Background

Surgical resection remains the gold standard for early-stage lung cancer. This study aimed to evaluate the predictive accuracy of the Kristersson/Olsen formula for postoperative lung function in patients undergoing minimally invasive surgery (MIS).

Methods

Seventy-five patients who underwent anatomical lung resection for lung cancer between January 2022 and January 2024 were prospectively analyzed. Preoperative, first postoperative week and second postoperative month pulmonary function test values (FEV₁, FVC), surgical approach, postoperative complications, and demographic and clinical variables were recorded.

Results

Regardless of surgical approach, the Kristersson/Olsen formula predicted a mean postoperative FEV₁ loss of 571 mL. However, the actual mean loss in the first postoperative week was significantly greater (698 mL; p = 0.013). In the MIS group, the predicted loss was 549 mL, whereas the observed first week loss reached 750 mL (p = 0.025). In contrast, during the second postoperative month, the overall patient cohort demonstrated a significantly smaller FEV₁ loss than predicted (449 mL vs. 571 mL; p = 0.002). Similarly, MIS patients showed an actual loss of 408 mL compared with the expected 549 mL (p = 0.013). When postoperative respiratory function recovery was examined, FEV₁ and FVC losses were significantly greater at the first postoperative week than at the second postoperative month in the MIS group, with a large effect size (p < 0.001, Cohen’s d 0.892 and 1.042, respectively). Thoracotomy patients also demonstrated greater early losses, with moderate-to-high effect sizes (Cohen’s d: 0.586 and 0.797). Recovery of pulmonary function from the first week to the second month was more pronounced in the MIS group.

Conclusions

The Kristersson/Olsen formula demonstrated limited accuracy in predicting postoperative pulmonary function during the early recovery phase (up to the second postoperative month), particularly in patients undergoing minimally invasive surgery. MIS is associated with lower-than-predicted functional loss, fewer complications, and greater pulmonary recovery, suggesting that patients with limited respiratory reserve may still be suitable candidates for surgery.

Background

According to the World Health Organization–International Agency for Research on Cancer (WHO-IARC) GLOBOCAN database, lung cancer ranks first among cancer-related causes of death worldwide [1]. Surgical resection is the gold standard treatment for early-stage lung cancer. In patients scheduled for lung resection, preoperative assessment of pulmonary function is one of the key determinants of postoperative morbidity and mortality [2]. In recent years, the prevalence of cancer patients with limited pulmonary reserve has increased, and a growing number of lung resections are being performed in this patient population.

Spirometry is the most commonly used and easily applicable method for assessing pulmonary function, providing standardized pulmonary function tests (PFTs). In patients scheduled for lung resection, PFTs constitute the cornerstone of preoperative evaluation. Several methods have been proposed in the literature to assess postoperative pulmonary function, including forced expiratory volume in 1 s (FEV₁), predicted postoperative forced expiratory volume in 1 s (ppoFEV₁), diffusion capacity of the lung for carbon monoxide (DLCO), and arterial blood gas analysis [3, 4]. The ppoFEV₁ value, calculated using the preoperatively measured FEV₁ and the number of lung segments to be resected, is a critical parameter for predicting whether patients will retain adequate pulmonary reserve following surgery.

Most of the assumptions regarding the predictive value of ppoFEV₁ measurements were established during an era when lung cancer surgery was predominantly performed using open surgical techniques. In contrast, there is limited evidence regarding the reliability and clinical applicability of this parameter following minimally invasive surgery (MIS). Contemporary surgical practice, supported by minimally invasive approaches, advanced analgesic strategies, and optimized perioperative care pathways such as Enhanced Recovery After Surgery (ERAS) protocols, has enabled surgical treatment even in patients with reduced cardiopulmonary reserve [5]. From this perspective, it can be hypothesized that postoperative FEV₁ values after MIS may be higher than the calculated ppoFEV₁. Accordingly, the present study aimed to evaluate the accuracy and effectiveness of the Kristersson/Olsen formula in predicting postoperative pulmonary function in patients undergoing MIS for lung cancer.

Methods

Patients who underwent anatomical lung resection (lobectomy, bilobectomy, pneumonectomy, or segmentectomy) for a diagnosis or suspected diagnosis of lung cancer in our clinic between January 2022 and January 2024 were considered for inclusion in the study. A total of 230 patients were initially assessed. After applying inclusion and exclusion criteria, complete data were available for 91 patients. During follow-up, patients who declined participation, were unable to comply with pulmonary function testing, did not undergo spirometry at the second postoperative month, were lost to follow-up, or had benign histopathology were excluded, resulting in a final study population of 75 patients. The patient selection process is summarized in Fig. 1. Demographic and clinical characteristics, including age and sex, surgical approach, postoperative complications, and pulmonary function test results (FEV₁ and FVC) obtained preoperatively, at the first postoperative week, and at the second postoperative month, were recorded and prospectively analyzed.

Fig. 1.

Patient flow diagram. Flowchart illustrating patient selection, application of inclusion and exclusion criteria, and final study population

Standardization of perioperative management and patient care

The MIS group consisted of patients who underwent VATS or hybrid surgical procedures. Although the definition of VATS lobectomy remains a subject of debate, patients were classified as having undergone VATS if the following criteria reported in the literature were met: a utility incision measuring 5–8 cm, the presence of a variable number of additional port incisions (2, 3, or 4) or a uniportal approach, avoidance of rigid rib retractors (with the optional use of soft wound protectors such as Alexis), and reliance on monitor-based visualization rather than direct visualization through the incision [6]. Patients operated on in accordance with these criteria were included in the VATS group. Patients who underwent surgery via a utility thoracotomy (< 10 cm) with the use of a rib retractor and thoracoscopic assistance (semi-open, semi-closed approach) were classified as the hybrid surgery group. Patients undergoing thoracotomy were included in the open surgery group regardless of whether a muscle-sparing or muscle-dividing technique was employed. In our clinic, instead of a standard posterolateral thoracotomy, a latissimus dorsi–sparing lateral thoracotomy or an anterolateral thoracotomy with an incision length of < 15 cm is generally preferred.

All surgical procedures were performed by surgeons trained and practicing at the same center, with comparable levels of experience. Anatomical lung resections (segmentectomy, lobectomy, bilobectomy, or pneumonectomy) with hilar and mediastinal lymph node dissection were carried out using either an MIS or thoracotomy approach, and stapling devices were used during all resections. Following resection, air-leak testing was performed under an airway pressure of 30 cmH₂O, and any detected leaks were repaired with sutures. With regard to perioperative analgesia, standard analgesic management consisted of opioid derivatives, paracetamol, and nonsteroidal anti-inflammatory drugs (NSAIDs). All patients were extubated in the operating room and transferred to the surgical ward for follow-up on the same day.

Postoperative pain control was achieved using intravenous (IV) opioids, IV paracetamol, and oral nonsteroidal anti-inflammatory drugs (NSAIDs). All patients were enrolled in a pulmonary rehabilitation program and received postoperative chest physiotherapy. Patients were evaluated daily with chest radiography, and thoracic drains were removed in the absence of air leakage or when fluid drainage was < 200 mL.

FEV₁ and FVC measurements and calculated values

Preoperative and postoperative spirometry-derived FEV₁ and FVC values, expressed in both milliliters and percentage of predicted, were recorded for all patients. Based on these measurements, predicted postoperative FEV₁ (ppoFEV₁, mL) and predicted postoperative FVC (ppoFVC, mL) values were calculated using the Kristersson/Olsen formula:

(where S represents the number of resected lung segments).

The difference between preoperative FVC and FEV₁ values and the calculated ppoFVC and ppoFEV₁ values was defined as the expected loss, whereas the difference between preoperative values and the measured postoperative values was considered the actual loss of FVC and FEV₁.

Pulmonary function measurements at the first postoperative week and the second postoperative month were performed by the same experienced physician and technician team using a portable spirometer (Spiromaster PC-10). All measurements were obtained with the patient in the seated position after a minimum rest period of 15 min. Patients were connected to the spirometer via a mouthpiece, and disposable mouthpieces were used for each measurement. To prevent nasal breathing, a nose clip was applied. During testing, patients were first instructed to breathe normally, during which vital capacity and resting respiratory status were assessed. Subsequently, patients were asked to take a deep inspiration followed by a maximal, rapid, and forceful expiration into the mouthpiece. Measured values were evaluated by comparison with predicted normal reference values adjusted for age, sex, height, and body weight. At least three measurements were obtained from each patient, and the best spirometric values were used for analysis. Device maintenance and calibration were performed in accordance with the European Respiratory Society (ERS) guidelines, with daily calibration using a 1000-mL calibration syringe.

Statistical analysis

Descriptive statistics were used to summarize continuous variables, including mean, standard deviation (SD), median, interquartile range (IQR), minimum, and maximum values, whereas categorical variables were analyzed using frequency distributions and percentages. One-way analysis of variance (ANOVA; Fisher’s F-test) was employed to assess differences among multiple groups. Independent-samples t-tests (Student’s t-test or Welch’s t-test, as appropriate) were used to compare mean differences in FEV₁ and FVC loss between independent groups. When the assumptions of normality were not met, the Kruskal–Wallis test was used for nonparametric comparisons among more than two independent groups. Paired-sample Wilcoxon signed-rank tests were applied to compare predicted and actual FEV₁/FVC losses within the same individuals at different postoperative time points (first week and second month). Effect sizes were calculated to quantify the magnitude of observed differences, using Cohen’s d for continuous variables and rank-biserial correlation for nonparametric comparisons. Statistical significance was set at an alpha level of 0.05, and p values < 0.05 were considered statistically significant. All analyses were conducted using RStudio (version 4.4.2), with Jamovi (version 2.6.23) employed for additional statistical verification. To account for baseline differences between surgical groups, multivariable linear regression analysis was performed. The model evaluated predictors of FEV₁ loss at the second postoperative month and included surgical approach, baseline FEV₁, age group, smoking status, and comorbidity status as covariates. A post-hoc power analysis was performed based on the observed effect size for differences in postoperative FEV₁ loss between surgical approaches.

Results

Patient characteristics and distribution according to study groups

The mean age of the patients was 65.07 ± 7.8 years (range, 46–79), and 74.7% were male. At least one comorbid condition was present in 56 patients (74.7%). Forty-five patients (60%) were current smokers, whereas 30 (40%) were never-smokers or former smokers. Forty patients (53.3%) underwent thoracotomy, and 35 (46.7%) were operated on using minimally invasive surgical approaches. Lobectomy was the most frequently performed procedure, accounting for 63 cases (84%), followed by pneumonectomy in 6 patients (8%), bilobectomy in 4 patients (5.3%), and segmentectomy in 2 patients (2.7%). Histopathological evaluation revealed adenocarcinoma in 35 patients (47%), squamous cell carcinoma in 22 (29%), large-cell neuroendocrine carcinoma in 6 (8%), carcinoid tumor in 5 (6.7%), metastatic lung disease in 5 (6.7%), small-cell lung carcinoma in 1 (1.3%), and pleomorphic carcinoma in 1 (1.3%). Postoperative complications occurred in 32 patients (42.7%). The most common complication was prolonged air leakage, observed in 14 patients (18.7%). Other complications included blood or blood-product transfusion in 7 patients (9.3%), cardiac complications in 5 (6.7%), subcutaneous emphysema in 2 (2.7%), chylothorax in 2 (2.7%), atelectasis/pneumonia in 1 (1.3%), and empyema in 1 (1.3%).

When the thoracotomy and MIS groups were compared in terms of baseline characteristics, no statistically significant differences were observed between the groups with respect to age, smoking history, comorbidities or type of surgical procedure. Male patients constituted 65% of the thoracotomy group and 85.7% of the MIS group, and this difference was statistically significant (p = 0.040). Postoperative complications occurred in 60% of patients in the thoracotomy group compared with 28.6% in the MIS group, representing a statistically significant difference (p = 0.006). The majority of patients in the MIS group (74.3%) were managed with a single chest drain, whereas 60% of patients in the thoracotomy group required two chest drains (p = 0.003). No statistically significant differences were found between the two surgical approaches in terms of time to chest drain removal or length of postoperative hospital stay (p = 0.80 and p = 0.081, respectively). However, the mean postoperative hospital stay was shorter in the MIS group compared with the thoracotomy group (5.3 ± 2 days vs. 7.2 ± 3 days). Demographic characteristics of the patients and group-specific features according to surgical approach are summarized in Table 1.

Table 1.

Demographic characteristics of patients and distribution according to surgical approach

Features	Total (N = 75)	MIS (N = 35)	Thoracotomy (N = 40)	p-value
Age group				0.090
<60	17 (23%)	11 (31%)	6 (15%)
≥60	58 (77%)	24 (69%)	34 (85%)
Sex				0.040
Male	56 (75%)	30 (86%)	26 (65%)
Female	19 (25%)	5 (14%)	14 (35%)
Smoking Group				0.600
Active	45 (60%)	19 (54%)	26 (65%)
Exsmoker	25 (33%)	13 (37%)	12 (30%)
Never smoked	5 (6.7%)	3 (8.6%)	2 (5.0%)
Comorbidity				> 0.900
Yes	56 (75%)	26 (74%)	30 (75%)
No	19 (25%)	9 (26%)	10 (25%)
Surgery				0.084
RUL	19 (25%)	9 (26%)	10 (25%)
RML	3 (4.0%)	3 (8.6%)	0 (0%)
RLL	11 (15%)	7 (20%)	4 (10%)
RUBL	2 (2.7%)	1 (2.9%)	1 (2.5%)
RLBL	2 (2.7%)	0 (0%)	2 (5.0%)
RP	1 (1.3%)	0 (0%)	1 (2.5%)
LUL	15 (20%)	5 (14%)	10 (25%)
LLL	15 (20%)	9 (26%)	6 (15%)
LP	5 (6.7%)	0 (0%)	5 (13%)
Segmentectomy	2 (2.7%)	1 (2.9%)	1 (2.5%)
Histopathology				0.200
Adenocarcinoma	35 (47%)	20 (57%)	15 (38%)
SCC	22 (29%)	7 (20%)	15 (38%)
Carcinoid	5 (6.7%)	3 (8.6%)	2 (5.0%)
Metastasis	5 (6.7%)	3 (8.6%)	2 (5.0%)
Pleomorphic carcinoma	1 (1.3%)	1 (2.9%)	0 (0%)
LCNEC	6 (8.0%)	1 (2.9%)	5 (13%)
SCLC	1 (1.3%)	0 (0%)	1 (2.5%)
Drains				0.003
1	42 (56%)	26 (74%)	16 (40%)
2	33 (44%)	9 (26%)	24 (60%)
Drain termination time (day)	5 ± 3, (Min:1, Max:17)	5 ± 4, (Min:2, Max:16)	5 ± 3, (Min:1, Max:17)	0.800
Hospital stay duration (day)	6 ± 3, (Min:3, Max:19)	5 ± 2, (Min:3, Max:11)	7 ± 3, (Min:3, Max:19)	0.081
Complication				0.006
Yes	34 (%45)	10 (%21)	24 (%60)
No	41 (%55)	25 (%79)	16 (%40)
Type of complication				0.020
Prolonged air leak	14 (18.7%)	2 (5.7%)	8 (20%)
Atelectasis/Pneumonia	1 (1.3%)	0 (0%)	1 (2.5%)
Empyema	1 (1.3%)	0 (0%)	1 (2.5%)
Chylothorax	2 (2.7%)	1 (2.9%)	1 (2.5%)
Subcutaneous emphysema	2 (2.7%)	1 (2.9%)	1 (2.5%)
Cardiac	5 (6.7%)	0 (0%)	5 (12.5%)
Blood/blood product replacement	7 (9.3%)	2 (5.7%)	5 (12.5%)
Other*	9 (12%)	2 (5.8%)	7 (17.5%)

Values are presented as n (%) or mean ± standard deviation (min–max). Pearson χ² test, Fisher’s exact test, or Wilcoxon rank-sum test were used as appropriate

Bold values indicate statistically significant results (p < 0.05)

MIS Minimally invasive surgery, RUL Right upper lobectomy, RML Right middle lobectomy, RLL Right lower lobectomy, RUBL Right upper bilobectomy, RLBL Right lower bilobectomy, RP Right pneumonectomy, LUL Left upper lobectomy, LLL Left lower lobectomy, LP Left pneumonectomy, SCC Squamous cell carcinoma, LCNEC Large cell neuroendocrine carcinoma, SCLC Small cell carcinoma, *Other; Glob vesicale, ileus, electrolyte imbalance, renal function test abnormalities, delirium

FVC and FEV₁ measurements

The mean preoperative FVC was 3297 ± 847 mL (range, 1460–5160 mL), and the mean preoperative FEV₁ was 2369 ± 720 mL (range, 1130–3920 mL). At the first postoperative week, the mean FVC decreased to 2225 ± 695 mL (range, 780–4630 mL), while the mean FEV₁ was 1672 ± 533 mL (range, 600–2970 mL). At the second postoperative month, mean FVC and FEV₁ values increased to 2628 ± 741 mL (range, 930–4580 mL) and 1920 ± 548 mL (range, 810–2950 mL), respectively. According to the Kristersson/Olsen formula, the calculated mean predicted postoperative values were 2518 ± 734 mL (range, 895–4393 mL) for ppoFVC and 1798 ± 596 mL (range, 626–3301 mL) for ppoFEV₁.

When preoperative pulmonary function values were compared between the groups, the thoracotomy group demonstrated a mean FVC of 3076 ± 850 mL (range, 1460–5160 mL) and a mean FEV₁ of 2152 ± 651 mL (range, 1130–3920 mL), whereas the MIS group showed higher baseline values, with a mean FVC of 3551 ± 779 mL (range, 1690–4910 mL) and a mean FEV₁ of 2617 ± 723 mL (range, 1160–3920 mL). These differences were statistically significant (p = 0.014 for FVC and p = 0.005 for FEV₁).

When postoperative pulmonary function measurements obtained at the first postoperative week were compared between the groups, the thoracotomy group had a mean FVC of 1991 ± 534 mL (range, 780–3070 mL) and a mean FEV₁ of 1500 ± 446 mL (range, 600–2540 mL). The MIS group demonstrated higher values, with a mean FVC of 2492 ± 766 mL (range, 855–2880 mL) and a mean FEV₁ of 1867 ± 562 mL (range, 780–2970 mL). These differences were statistically significant (p = 0.002 for FVC and p = 0.003 for FEV₁).

When pulmonary function measurements obtained at the second postoperative month were compared between the groups, the thoracotomy group demonstrated a mean FVC of 2351 ± 692 mL (range, 930–4130 mL) and a mean FEV₁ of 1668 ± 465 mL (range, 810–2840 mL). The MIS group showed higher values, with a mean FVC of 2945 ± 671 mL (range, 1170–4580 mL) and a mean FEV₁ of 2209 ± 494 mL (range, 890–2950 mL). The differences between the groups were statistically significant (p < 0.001). A comparative summary of FVC and FEV₁ measurements is presented in Table 2.

Table 2.

Actual FVC and FEV₁ measurements in the overall patient cohort and stratified by surgical approach

Features	Total (N = 75)	MIS (N = 35)	Thoracotomy (N = 40)	p-value
Preop FVC (mL)	3,297 ± 847, (Min:1,460, Max:5,160)	3,551 ± 779, (Min:1,690, Max:4,910)	3,076 ± 850, (Min:1,460, Max:5,160)	0.014
Preop FVC (%)	98 ± 16, (Min:61, Max:145)	100 ± 16, (Min:61, Max:127)	97 ± 17, (Min:68, Max:145)	0.400
Preop FEV1 (mL)	2,369 ± 720, (Min:1,130, Max:3,920)	2,617 ± 723, (Min:1,160, Max:3,920)	2,152 ± 651, (Min:1,130, Max:3,920)	0.005
Preop FEV1 (%)	89 ± 20, (Min:42, Max:138)	93 ± 20, (Min:42, Max:124)	86 ± 19, (Min:46, Max:138)	0.120
Ppo FVC (mL)	2,518 ± 734, (Min:895, Max:4,393)	2,820 ± 655, (Min:1,334, Max:4,393)	2,253 ± 703, (Min:895, Max:4,345)	< 0.001
Ppo FVC (%)	75 ± 17, (Min:44, Max:122)	80 ± 15, (Min:45, Max:112)	72 ± 18, (Min:44, Max:122)	0.040
Ppo FEV1 (mL)	1,798 ± 596, (Min:626, Max:3,301)	2,069 ± 563, (Min:855, Max:2,888)	1,561 ± 524, (Min:626, Max:3,301)	< 0.001
Ppo FEV1 (%)	68 ± 18, (Min:31, Max:116)	74 ± 17, (Min:31, Max:104)	63 ± 18, (Min:34, Max:116)	0.011
1st week FVC (mL)	2,225 ± 695, (Min:780, Max:4,630)	2,492 ± 766, (Min:1,070, Max:4,630)	1,991 ± 534, (Min:780, Max:3,070)	0.002
1st week FVC (%)	67 ± 18, (Min:35, Max:111)	72 ± 19, (Min:38, Max:110)	63 ± 17, (Min:35, Max:111)	0.043
1st week FEV1 (mL)	1,672 ± 533, (Min:600, Max:2,970)	1,867 ± 562, (Min:780, Max:2,970)	1,500 ± 446, (Min:600, Max:2,540)	0.003
1st week FEV1 (%)	63 ± 18, (Min:29, Max:104)	67 ± 19, (Min:29, Max:104)	59 ± 17, (Min:33, Max:104)	0.057
2nd month FVC (mL)	2,628 ± 741, (Min:930, Max:4,580)	2,945 ± 671, (Min:1,170, Max:4,580)	2,351 ± 692, (Min:930, Max:4,130)	< 0.001
2nd month FVC (%)	79 ± 19, (Min:42, Max:125)	85 ± 16, (Min: 45, Max: 120)	74 ± 21, (Min: 42, Max: 125)
2nd month FEV1 (mL)	1,920 ± 548, (Min:810, Max:2,950)	2,209 ± 494, (Min:890, Max:2,950)	1,668 ± 465, (Min:810, Max:2,840)	< 0.001
2nd month FEV1 (%)	72 ± 18, (Min:41, Max:112)	80 ± 16, (Min:41, Max:112)	66 ± 18, (Min:45, Max:110)

Values are presented as mean ± standard deviation (min–max). Group comparisons were performed using the Welch two-sample t-test

Bold values indicate statistically significant results (p < 0.05)

MIS: minimally invasive surgery

Evaluation of predicted and actual FVC and FEV₁ loss

According to the Kristersson/Olsen formula, the mean predicted FVC loss was 780 ± 364 mL (range, 154–2253 mL), and the mean predicted FEV₁ loss was 571 ± 291 mL (range, 123–1711 mL). At the first postoperative week, the mean actual FVC loss was 1072 ± 640 mL (range, − 90 to 2400 mL), and the mean actual FEV₁ loss was 698 ± 543 mL (range, − 380 to 2240 mL). At this time point, the actual FVC loss was significantly greater than the predicted FVC loss (p < 0.001). Similarly, the actual FEV₁ loss at the first postoperative week exceeded the predicted FEV₁ loss, and this difference was statistically significant (p = 0.013). At the second postoperative month, the mean actual FVC loss was 669 ± 550 mL (range, − 430 to 2190 mL), and the mean actual FEV₁ loss was 449 ± 417 mL (range, − 590 to 1600 mL). Postoperative FEV₁ values exceeding baseline values were observed in 4 patients (5.3%) at postoperative week 1 and in 7 patients (9.3%) at the second postoperative month. In contrast to the first postoperative week, the actual FVC loss at the second month was significantly lower than the predicted FVC loss (p = 0.036). Likewise, the actual FEV₁ loss was significantly lower than the predicted FEV₁ loss (p = 0.002) (Fig. 2).

Fig. 2.

Overall comparison of predicted and actual FVC and FEV₁ loss (mL)

In the subsequent analysis, patients who underwent thoracotomy and MIS were evaluated separately within their respective groups (Fig. 3). In the MIS group, the mean predicted FVC loss calculated using the Kristersson/Olsen formula was 731 ± 265 mL (range, 334–1216 mL), and the mean predicted FEV₁ loss was 549 ± 243 mL (range, 224–1032 mL). At the first postoperative week, the mean actual FVC loss in the MIS group was 1059 ± 644 mL (range, − 90 to 2400 mL), and the mean actual FEV₁ loss was 750 ± 653 mL (range, − 380 to 2240 mL). At this time point, the actual FVC loss was significantly greater than the predicted FVC loss (p < 0.001). Similarly, the actual FEV₁ loss at the first postoperative week exceeded the predicted FEV₁ loss, and this difference was statistically significant (p = 0.025). At the second postoperative month, the mean actual FVC loss in the MIS group was 605 ± 660 mL (range, − 430 to 1510 mL), and the mean actual FEV₁ loss was 408 ± 417 mL (range, − 590 to 1110 mL). At this time point, the difference between predicted and actual postoperative FVC loss was not statistically significant (p = 0.070). In contrast, the actual FEV₁ loss at the second postoperative month was significantly lower than the predicted FEV₁ loss (p = 0.013). In patients undergoing open surgery, the mean predicted FVC loss was calculated as 823 ± 431 mL (range, 154–2253 mL), and the mean predicted FEV₁ loss was 591 ± 329 mL (range, 123–1711 mL). At the first postoperative week, the mean actual FVC loss was 1084 ± 1030 mL (range, 40–2390 mL), and the mean actual FEV₁ loss was 652 ± 428 mL (range, − 230 to 1720 mL). At this time point, the actual FVC loss was significantly greater than the predicted FVC loss (p = 0.005), whereas no statistically significant difference was observed between actual and predicted FEV₁ loss (p = 0.264). At the second postoperative month, patients in the thoracotomy group demonstrated a mean actual FVC loss of 725 ± 603 mL (range, − 80 to 2190 mL) and a mean actual FEV₁ loss of 485 ± 401 mL (range, − 80 to 1600 mL). No statistically significant difference was observed between predicted and actual postoperative FVC loss at this time point (p = 0.220). In contrast, the difference between predicted and actual postoperative FEV₁ loss at the second month reached statistical significance, with a small positive effect size (p = 0.047). The results of analyses comparing predicted and actual FVC and FEV₁ losses are summarized in Table 3.

Fig. 3.

Comparison of predicted and actual FEV₁ loss (mL) according to surgical approach

Table 3.

Comparison of predicted and actual FVC and FEV₁ loss

Features	N	Mean (Min-Max) (mL)	Median (mL)	SD (mL)	SE (mL)	p-value	Mean difference (mL)	Effect size (Cohen’s d)
All Groups
Predicted FVC loss	75	780(154–2253)	742	364	42.0	-	-	-
Actual FVC loss (1st week)	75	1072(-90-2400)	920	640	73.9	< 0.001	-293	-0.542
Actual FVC loss (2nd month)	75	669(-430-2190)	550	550	63.5	0.036	111	0.246
Predicted FEV1 loss	75	571(123–1711)	500	291	33.6	-	-	-
Actual FEV1 loss (1st week)	75	698(-380-2240)	650	543	62.7	0.013	-126	-0.295
Actual FEV1 loss (2nd month)	75	449(-590-1600)	400	417	48.2	0.002	122	0.380
MIS group
Predicted FVC loss	35	731(334–1216)	707	265	44.7	-	-	-
Actual FVC loss (1st week)	35	1059(-90-2400)	879	644	108.9	< 0.001	-328	-0.627
Actual FVC loss (2nd month)	35	605(-430-1510)	650	660	87.8	0.070	125	0.316
Predicted FEV1 loss	35	549(224–1032)	493	243	41.1	-	-	-
Actual FEV1 loss (1st week)	35	750(-380-2240)	650	653	110.4	0.025	-201	-0.398
Actual FEV1 loss (2nd month)	35	408(-590-1110)	404	417	73.9	0.013	141	0.442
Thoracotomy group
Predicted FVC loss	40	823(154–2253)	784	431	68.1	-	-	-
Actual FVC loss (1st week)	40	1084(40-2390)	1030	644	101.8	0.005	-261.7	-0.468
Actual FVC loss (2nd month)	40	725(-80-2190)	560	603	95.3	0.220	97.8	0.197
Predicted FEV1 loss	40	591(123–1711)	525	329	52.0	-	-	-
Actual FEV1 loss (1st week)	40	652(-230-1720)	620	428	67.6	0.264	-60.9	-0.179
Actual FEV1 loss (2nd month)	40	485(-80-1600)	401	401	63.5	0.047	106.4	0.324

Bold values indicate statistically significant results (p < 0.05)

SD standard deviation, SE standard error, df degrees of freedom, MIS Minimally invasive surgery

Multivariable analysis of predictors of postoperative FEV₁ loss

Multivariable linear regression analysis adjusting for baseline FEV₁, age group, smoking status, and comorbidities demonstrated that thoracotomy remained independently associated with greater postoperative FEV₁ loss at the second postoperative month compared with minimally invasive surgery (β = 323.1 mL, p < 0.001). Baseline FEV₁ was also significantly associated with postoperative FEV₁ loss (β = 0.418, p < 0.001), whereas age group, smoking status, and comorbidity status were not significant predictors (Table 4).

Table 4.

Multivariable linear regression analysis of predictors of FEV₁ loss at the second postoperative month

Variable	β coefficient (mL)	95% CI	p-value
Thoracotomy vs. MIS	323.1	(160.0–486.1)	< 0.001
Baseline FEV₁	0.418	(0.298–0.538)	< 0.001
Age ≥ 60 years	-199.4	(-380.7 – -18.1)	0.10
Smoking status	2.11	(-123.3–127.5)	0.55
Comorbidity	-11.6	(83.6–60.4)	0.41

Outcome variable: FEV₁ loss at the second postoperative month

Bold values indicate statistically significant results (p < 0.05)

MIS minimally invasive surgery. F Model R² = 0.56; adjusted R² = 0.52

These findings indicate that the observed differences between surgical approaches were not solely explained by baseline pulmonary function differences. Thoracotomy remained independently associated with greater FEV₁ loss at the second postoperative month compared with minimally invasive surgery (β = 323.1 mL, p < 0.001). The regression model demonstrated good explanatory performance (R² = 0.56; adjusted R² = 0.52). Based on the observed effect size (Cohen’s d = 0.71), the post-hoc statistical power for detecting differences in postoperative FEV₁ loss between groups was 0.86 at an alpha level of 0.05.

Temporal changes in actual FVC and FEV₁ loss

When actual FVC and FEV₁ losses at the first postoperative week and the second postoperative month were compared in patients undergoing MIS, losses at the first week were found to be significantly greater than those observed at the second month (p < 0.001). The corresponding effect sizes (Cohen’s d) indicated a large magnitude of change for both FVC and FEV₁ (0.892 and 1.042, respectively).

Similarly, in patients who underwent thoracotomy, actual FVC and FEV₁ losses at the first postoperative week were significantly greater than those at the second postoperative month (p < 0.001). The effect sizes (Cohen’s d) were of moderate-to-high magnitude (0.586 for FVC and 0.797 for FEV₁) (Table 5).

Table 5.

Temporal comparison of actual FVC and FEV₁ loss

Group	Comparison	Test	t	df	p-value	Effect size (Cohen’s d)
MIS	FVC (1st week) - FVC (2nd month)	Student’s	6.17	34	< 0.001	1.042
	FEV1 (1st week) - FEV1 (2nd month)	Student’s	5.28	34	< 0.001	0.892
Thoracotomy	FVC (1st week) - FVC (2nd month)	Student’s	5.04	39	< 0.001	0.797
	FEV1 (1st week) - FEV1 (2nd month)	Student’s	3.71	39		0.586

Bold values indicate statistically significant results (p < 0.05)

df degrees of freedom, t distribution, MIS Minimally invasive surgery

Association between FVC and FEV₁ loss and postoperative complications

Postoperative complications occurred in 42.7% of the patients. When the relationship between FVC and FEV₁ loss and the presence of complications was evaluated, no statistically significant differences were observed between patients with and without complications in terms of first week or second month FVC loss (p = 0.612 and p = 0.683, respectively). Similarly, no statistically significant differences were found between first week and second month FEV₁ loss and the presence of postoperative complications (p = 0.414 and p = 0.367, respectively) (Table 6).

Table 6.

Association between first week and second month FVC and FEV₁ loss and postoperative complications

Category	Variable	N	Mean (mL)	Median (mL)	SD	SE	p-value
Complication status	FVC loss (1st week)						0.612
	No	41	1037	879	582	90.9
	Yes	34	1115	1105	710	121.8
	FEV1 loss (1st week)						0.414
	No	41	745	650	514	80.3
	Yes	34	640	620	578	99.1
	FEV1 loss (2nd month)						0.367
	No	41	491	450	300	46.9
	Yes	34	398	375	526	90.2
	FVC loss (2nd month)						0.683
	No	41	644	610	414	64.6
	Yes	34	699	565	684	117.4

SD standard deviation, SE standard error

Discussion

One of the most important causes of morbidity and mortality in lung cancer surgery is the inability of the remaining lung parenchyma to provide adequate respiratory function, potentially leading to postoperative respiratory failure. Consequently, limited pulmonary function test (PFT) results represent one of the major contraindications to lung cancer surgery. As the majority of patients with lung cancer are smokers, concomitant cardiovascular diseases and pulmonary parenchymal disorders such as chronic obstructive pulmonary disease (COPD) are frequently encountered in this population. These comorbid conditions contribute to impaired cardiopulmonary reserve [7]. For this reason, thorough preoperative evaluation is of critical importance in patients scheduled for lung resection. In our study, considering the mean patient age of 65 years and the presence of a smoking history in 93.3% of the cohort, the study population can be regarded as being at high risk for postoperative morbidity, in line with findings reported in the existing literature.

Even among patients with early-stage disease, there exists a subset of individuals who are considered medically inoperable. Medical inoperability is most often attributed to significant comorbidities and limited pulmonary reserve. Such decisions are typically based on algorithms and predictive formulas used during preoperative assessment. However, these algorithms were originally developed in an era when thoracic surgery was performed predominantly via open thoracotomy, with limited perioperative and postoperative analgesic options. With the continuous evolution of thoracic surgical techniques, we believe that these preoperative assessment algorithms should also be re-evaluated and updated. Undoubtedly, while the original criteria were established through extensive and long-term studies, their refinement and validation in the context of modern surgical practice will likewise require robust and methodologically sound investigations.

The Kristersson/Olsen formula is one of the most commonly used tools for predicting postoperative pulmonary reserve. Zeiher et al. [8] demonstrated a strong correlation between predicted postoperative FEV₁ (ppoFEV₁) calculated using the Kristersson/Olsen formula and measured postoperative values, particularly at the third and sixth postoperative months in patients undergoing lobectomy. In contrast, studies by Valera et al. [9, 10] reported a lack of correlation between ppoFEV₁ and actual FEV₁ in the early postoperative period, concluding that ppoFEV₁ is an inadequate predictor of early postoperative pulmonary reserve. Given that many postoperative complications occur during the early postoperative phase, these findings further suggest that ppoFEV₁ may be insufficient for predicting such complications. Our results are consistent with these observations. When the entire cohort was evaluated irrespective of surgical approach, the actual FEV₁ loss measured in the first postoperative week was significantly greater than the expected loss calculated using the Kristersson/Olsen formula (698 mL vs. 571 mL; p = 0.013). Similarly, actual FVC loss in the first week exceeded the predicted loss (1072 mL vs. 780 mL; p < 0.001). In contrast, during the second postoperative month, actual FEV₁ and FVC losses were significantly lower than the predicted values (571 mL vs. 449 mL for FEV₁, p = 0.002; and 780 mL vs. 669 mL for FVC, p = 0.036). Comparable findings were also observed when patients undergoing MIS and thoracotomy were analyzed separately. Even among patients who underwent thoracotomy, results similar to those in the MIS group were obtained. We speculate that this may be attributable to the use of smaller, muscle-sparing thoracotomy incisions rather than conventional posterolateral thoracotomy. Indeed, several studies have demonstrated that muscle-sparing lateral or anterolateral thoracotomy is associated with lower morbidity compared with classical posterolateral thoracotomy [11, 12]. Collectively, these findings suggest that ppoFEV₁ values may not fully reflect postoperative pulmonary function during the early recovery phase following lung resection. Moreover, the observation that the second postoperative month actual FEV₁ values exceed predicted ppoFEV₁ values suggests that surgical treatment may be feasible in a broader group of patients who were previously considered unsuitable for surgery. In a small subset of patients, postoperative spirometric values exceeded preoperative measurements, resulting in negative calculated loss values. This phenomenon has been described in previous studies and may reflect relief of airflow obstruction after resection of poorly ventilated lung segments, as well as variability in patient effort during spirometric testing [8, 9].

In addition to segment-based predictive formulas such as the Kristersson/Olsen method, several more advanced approaches have been proposed to estimate postoperative pulmonary function. These include CT volumetry, perfusion scintigraphy, and quantitative CT-based segmental analysis, which may provide more individualized assessments of postoperative functional capacity. However, these techniques require specialized imaging analysis and are not routinely available in all clinical settings. In contrast, the Kristersson/Olsen formula remains widely used in daily clinical practice because of its simplicity, accessibility, and ability to provide a rapid estimation of postoperative pulmonary reserve during preoperative evaluation. Recent studies have demonstrated that CT volumetry and quantitative CT-based segmental analysis may provide more individualized estimates of postoperative pulmonary function by accounting for regional lung volume and parenchymal quality. Similarly, perfusion scintigraphy may improve prediction accuracy by evaluating the functional distribution of pulmonary perfusion between lung segments. These techniques may therefore complement traditional segment-based formulas such as the Kristersson/Olsen method, particularly in patients with heterogeneous lung disease. Quantitative CT volumetry has been shown to improve the accuracy of postoperative FEV₁ prediction by accounting for regional lung volume and parenchymal heterogeneity, particularly in patients with underlying emphysematous changes [13, 14]. Quantitative CT-based analyses have also demonstrated the ability to evaluate postoperative compensatory changes in the residual lung and to provide a more individualized assessment of functional recovery after lobectomy and segmentectomy [15]. More recently, artificial intelligence–assisted predictive models integrating imaging, clinical, and functional variables have demonstrated promising performance in postoperative risk stratification and prediction of pulmonary complications [16]. Although these techniques are not yet routinely implemented in daily clinical practice and require specialized imaging analysis, they may provide a more individualized estimation of postoperative pulmonary reserve than conventional segment-counting methods. Furthermore, the improved preservation of chest wall integrity and respiratory mechanics associated with minimally invasive surgical approaches may reduce the reliance on purely segment-based predictions, as functional recovery may be influenced not only by the extent of resection but also by perioperative factors affecting lung mechanics [17]. Future studies integrating imaging-based predictive tools with perioperative and surgical variables may further refine functional risk stratification in contemporary thoracic surgery.

When temporal changes in pulmonary function were examined, significant differences were observed between first week and second month postoperative losses in both the MIS and thoracotomy groups (p < 0.001). However, recovery of FVC and FEV₁ from the first postoperative week to the second postoperative month was more pronounced in patients undergoing MIS compared with those who underwent thoracotomy, with a strong and clinically meaningful magnitude of improvement (Cohen’s d: 0.892–1.042). Previous studies have emphasized that MIS is associated with reduced postoperative pain and less functional impairment due to smaller incisions and muscle-sparing techniques [18–20]. Our findings are consistent with these observations and further demonstrate that recovery of lost pulmonary function is greater in patients undergoing MIS than in those treated with thoracotomy.

Nakahara et al. [21] demonstrated a negative correlation between FEV₁ and postoperative complications in a cohort of 188 patients undergoing anatomical lung resection for lung cancer, reporting a significantly higher incidence of complications in patients with a predicted %FEV₁ of less than 60% of normal (p < 0.05). The authors emphasized that ppoFEV₁ calculated using the Kristersson/Olsen formula may be useful for predicting postoperative complications, particularly in elderly patients. In contrast, in our study, although first week and second month postoperative FEV₁ and FVC losses showed slight differences between patients with and without complications, these differences did not reach statistical significance. The absence of statistically significant associations between pulmonary function loss and postoperative complications should be interpreted with caution, as the study was not specifically powered to detect clinically meaningful differences in complication rates between groups and may also reflect limited statistical power for detecting small effect sizes in this cohort. Accordingly, these findings should not be interpreted as evidence that postoperative pulmonary function loss is unrelated to complications, but rather as an indication that larger studies are needed to clarify the magnitude and clinical relevance of this association.

In contemporary lung cancer surgery, MIS has been shown to be associated with less postoperative pain, shorter hospital stay, lower complication rates, and improved survival compared with thoracotomy, and in recent years has even been advocated as the gold standard surgical approach [22–27]. Consistent with the existing literature, our study demonstrated a significantly lower rate of postoperative complications in patients undergoing MIS compared with those treated via thoracotomy (29% vs. 60%, p = 0.006). When patients were evaluated in terms of time to chest drain removal and length of postoperative hospital stay, these durations were shorter in the MIS group; however, the differences did not reach statistical significance (p = 0.800 and p = 0.081, respectively). Notably, the mean postoperative length of stay was 5 ± 2 days in the MIS group compared with 7 ± 3 days in the thoracotomy group, with the difference approaching statistical significance. These findings suggest that studies with larger patient populations may yield results more closely aligned with those reported in the literature.

Strengths of the study

This study has several strengths. First, the data were collected prospectively, allowing standardized assessment of pulmonary function at predefined postoperative time points. Second, the comparison between minimally invasive surgery and thoracotomy was supported by multivariable regression analysis adjusting for baseline pulmonary function and key clinical variables, thereby reducing potential confounding. Third, both predicted and measured pulmonary function parameters were evaluated, providing a comprehensive assessment of the predictive performance of the Kristersson/Olsen formula in contemporary thoracic surgical practice. Fourth, pulmonary function measurements were obtained using standardized spirometry performed by the same experienced team.

Limitations

This study has several limitations that should be acknowledged. First, it was conducted at a single center with a relatively limited sample size, which may restrict the generalizability of the findings. Although the study population was homogeneous in terms of surgical expertise and perioperative management, multicenter studies with larger cohorts are required to validate our results. Second, despite prospective data collection, the study design was observational rather than randomized. Therefore, potential selection bias cannot be entirely excluded, particularly in the allocation of patients to minimally invasive surgery or thoracotomy, which was based on clinical judgment. In addition, baseline pulmonary function values were significantly higher in the MIS group, which may have influenced postoperative functional recovery despite the adjusted analyses performed. Third, postoperative pulmonary function was evaluated only up to the second postoperative month. Because pulmonary recovery may continue for several months after lung resection, longer-term follow-up would be necessary to fully assess the predictive accuracy of ppoFEV₁ for long-term functional outcomes. Finally, although postoperative complications were systematically recorded, the study was not specifically powered to detect subtle associations between pulmonary function changes and individual complication subtypes. Future studies incorporating larger patient populations and longer follow-up periods may better clarify these relationships.

Conclusion

The widespread adoption of minimally invasive surgical techniques has had a substantial impact on postoperative complications and pulmonary function outcomes. In addition to advances in surgical approaches, optimization of perioperative care through protocols such as Enhanced Recovery After Surgery (ERAS) has enabled lung resections to be performed with fewer complications and less functional impairment. The Kristersson/Olsen formula demonstrated limited accuracy in predicting early postoperative pulmonary function in this cohort, particularly among patients undergoing minimally invasive surgery. During the second postoperative month, patients undergoing MIS demonstrated a smaller than predicted loss of FEV₁ and a greater degree of pulmonary function recovery compared with those treated via thoracotomy. However, longer-term follow-up studies are required to determine its predictive value for long-term functional outcomes. These findings suggest that some patients previously considered high risk because of limited pulmonary function may still be suitable candidates for lung resection when modern surgical techniques are applied. Consequently, a broader patient population may benefit from surgical treatment. The ongoing evolution of surgical techniques and preoperative rehabilitation programs underscores the need for updated preoperative assessment guidelines that reflect contemporary thoracic surgical practice.

Abbreviations

DLCO

Diffusion capacity of the lung for carbon monoxide

ERAS

Enhanced recovery after surgery

ERS

European Respiratory Society

FEV₁

Forced expiratory volume in 1 second

FVC

Forced vital capacity

GGO

Ground glass opacity

IARC

International Agency for Research on Cancer

IQR

Interquartile range

SCLC

Small cell lung cancer/small cell carcinoma

NSCLC

Non–small cell lung cancer

COPD

Chronic obstructive pulmonary disease

LCNEC

Large cell neuroendocrine carcinoma

SCC

Squamous cell carcinoma

AC

Adenocarcinoma

MIS

Minimally invasive surgery

NSAIDs

Nonsteroidal anti-inflammatory drugs

PpoFEV₁

Predicted postoperative forced expiratory volume in 1 s

PpoFVC

Predicted postoperative forced vital capacity

PFT

Pulmonary Function Tests

VATS

Video-assisted thoracic surgery

WHO

World Health Organization

SD

Standard Deviation

Authors’ contributions

HEB; Conception and design, Collection and assembly of data, Data analysis and interpretation, Manuscript writing, Final approval of manuscript. TIA; Conception and design, Provision of study materials or patients, Manuscript writing, Revision, Final approval of manuscript. AKT; Conception and design, Collection and assembly of data, Data analysis and interpretation, Manuscript writing, Final approval of manuscript. GKO; Conception and design, Collection and assembly of data, Data analysis and interpretation, Manuscript writing, Final approval of manuscript. AGE; Provision of study materials or patients, Manuscript writing, Revision, Final approval of manuscript. AO; Provision of study materials or patients, Manuscript writing, Revision, Final approval of manuscript. AC; Provision of study materials or patients, Manuscript writing, Revision, Final approval of manuscript. UC; Provision of study materials or patients, Manuscript writing, Revision, Final approval of manuscript.

Funding

This work was financially supported by the Ege University Scientific Research Projects Coordination Unit (BAP) (Project No: 25219).

Data availability

The dataset generated and analyzed during the current study is available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

All procedures performed in this research were in compliance with the ethical standards of the institutional and national research committees and with the Declaration of Helsinki. Written informed consent was obtained from all participants or their legal guardians. Ethical approval for this study was obtained from the Ege University Faculty of Medicine Clinical Research Ethics Committee with decision number 22 − 1/53 dated January 11, 2022.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.