3DCT reconstruction—does 3DCT improve anatomical lung resection?—a narrative review of the literature Authors: Ilecia Baboolal, Kelvin Lau, Jose Alvarez Gallesio, Steven Stamenkovic, Tim J P Batchelor Published: 2025 ABSTRACT AbstractBackground and ObjectiveMinimally invasive and sublobar anatomical lung resection has become the gold standard for stage I non-small cell lung cancer (NSCLC) under 2 cm. Preoperative surgical planning with three-dimensional computed tomography (3DCT) reconstruction has become more common to improve safety and accuracy of resection. The aim of this review is to evaluate the evidence for 3DCT reconstruction in improving perioperative outcomes in patients undergoing anatomical lung resection.MethodsA targeted literature review of evidence for 3DCT reconstruction in the perioperative period, focusing on quantitative data. Studies were included if they offered comparative data between two-dimensional computed tomography (2DCT) and 3DCT, or if they reported outcomes directly influenced using 3D imaging. Articles were excluded if they did not address preoperative imaging strategies, lacked peer review, or failed to provide sufficient data for analysis.Key Content and FindingsForty papers were identified for this review. Seventeen only described bronchovascular patterns and anatomical variations with no surgical procedure performed and were excluded. Twenty-three described 3DCT reconstruction in relation to surgical resection. Nine studies assessed resection margins, with one reporting a change from segmentectomy to lobectomy due to 3DCT findings (10.5%). Two showed improved margin adequacy with 3DCT, though overall evidence remains limited. Across six comparative studies, two reported reduced blood loss with 3DCT, while others showed no difference. Operative time showed mixed results: three retrospective studies reported shorter durations, though the single randomised controlled trial, the DRIVATS study, found no difference. There was also no difference in clinical outcomes such as chest tube drainage, postoperative complications and postoperative hospital stay. However, this may be due to segmentectomy being a heterogeneous group of operations, as well as underpowered studies.ConclusionsAlthough randomized evidence demonstrating the superiority of 3DCT reconstruction over conventional computed tomography (CT) is lacking, 3DCT remains a valuable adjunct for visualising complex anatomical structures and guiding operative planning.Keywords: Minimally invasive thoracic surgery, three-dimensional computed tomography reconstruction (3DCT reconstruction), sublobar lung resection Introduction Two large, randomised control trials, CALGB 140503 and JCOG 0802, established the role of sublobar resection in the treatment of early-stage lung cancer under 2 cm and recent guidelines have been updated to recognise this as the standard of care (1-3). Anatomical sublobar resections are more complex due to large variations in segmental and subsegmental anatomy. A lack of clear anatomical boundaries between segments increases the chance of misidentification of the nodule’s segment, misidentification of the intersegmental plane, and insufficient margin. The need to better visualize and understand complex anatomical relationships preoperatively has become increasingly important. Surgeons traditionally rely on conventional CT scans visualised as consecutive two-dimensional (2D) slices to localize pulmonary lesions and decipher their relationship to surrounding anatomical structures. To interpolate what needs to be done at operation, the surgeon needs to compute a three-dimensional (3D) view based on this 2D information. This is a challenging skill with many pitfalls. Consequently, three-dimensional computed tomography (3DCT) reconstruction, providing a 3D view that can be manipulated, is increasingly becoming a routine part of the preoperative preparation to facilitate precise surgery. The current clinical applications of 3DCT in anatomical lung resection involve: preoperative mapping of the bronchovascular anatomy and variants, of which the aim is to reduce the chance of unexpected findings intraoperatively and avoid mishaps (4). Thoracic surgeons are using 3DCT more for segment localization and margin planning as we move to smaller anatomical sub-segmentectomies, while maintaining complete resection (5). It also provides the opportunity for surgeons to have a clearer stepwise surgical plan with a better understanding of the anatomy, even predicting difficult resections, which facilitates its usage as a training adjunct (6). There are multiple studies that seek to validate these potential uses of 3DCT reconstruction. We aim to review the objective evidence of these clinical applications, in the form of measurable outcomes such as operative time, blood loss, resection margin etc. We present this article in accordance with the Narrative Review reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1836/rc). Methods A targeted literature review was conducted using the PubMed database on February 8, 2025, to evaluate the role of 3DCT reconstruction compared to conventional 2DCT imaging in the preoperative planning of pulmonary resection. The search strategy included key terms related to 3D reconstruction, minimally invasive thoracic surgery, and specific types of pulmonary resection such as lobectomy and segmentectomy. The complete search string is provided in Appendix 1. Studies were selected based on their relevance to the use of 3DCT in preoperative planning and were included if they offered comparative data between 2DCT and 3DCT, or if they reported outcomes directly influenced by the use of 3D imaging. Articles were excluded if they did not address preoperative imaging strategies, lacked peer review or failed to provide sufficient data for analysis. Other articles not included were case reports and review papers, as well as those that looked at 3D printing or artificial intelligence (AI) exclusively. Due to our analysis of anatomical lung resection, we did not include studies that looked at mediastinal or chest wall tumours, pancoast tumours and bullae resection. We also excluded nodule analysis using 3DCT reconstruction and those looking in-depth into software development. The selection process, including inclusion and exclusion criteria, is detailed in the flowchart presented in Figure 1 and Table 1. Flowchart of search and selection process. 3D, three-dimensional; AI, artificial intelligence. Table 1The search strategy summaryItemsSpecificationDate of search8 February 2025Database searchedPubMed Search terms used3D imaging OR 3-D imaging OR 3 dimensional imaging OR three dimensional imaging OR 3-dimensional imaging OR three-dimensional imaging AND pulmonary resection OR lung resection OR thoracic surgery OR pulmonary surgery OR VATS OR video assisted thoracic surgery OR video assisted thorascopic surgery OR RATS OR robotic assisted thoracic surgery OR robotic assisted thorascopic surgery OR lobectomy OR segmentectomy OR sub segmentectomy OR sub-segmentectomy OR sleeve resection OR pneumonectomy OR pneumectomy OR wedge resectionTimeframe2008–2025Inclusion and exclusion criteriaInclusion: any study which used 3DCT in preoperative planning, if they offered comparative data between 2DCT and 3DCT. If they reported outcomes directly influenced using 3D imagingExclusion: case reports, review papers, software development studies, studies with 3D printing or AI only, mediastinal or chest wall tumours, pancoast tumours, resection of lung bullae only, studies on nodule characteristicsSelection processMs I.B. conducted the selection (corresponding author). This was conducted independently2DCT, two-dimensional computed tomography; 3D, three-dimensional; 3DCT, three-dimensional computed tomography; AI, artificial intelligence. The search strategy summary 2DCT, two-dimensional computed tomography; 3D, three-dimensional; 3DCT, three-dimensional computed tomography; AI, artificial intelligence. In evaluating the clinical impact of 3DCT reconstruction, the review focused on three key outcome domains that collectively reflect its utility in thoracic surgical planning. Pathological outcome, which assesses resection margins, as the adequacy of margin clearance correlates with local recurrence rates and overall survival (7). We also assessed missed nodules and completion of resection with R0 (resection 0—microscopically clear margins) clearance. Pathological outcome, which assesses resection margins, as the adequacy of margin clearance correlates with local recurrence rates and overall survival (7). We also assessed missed nodules and completion of resection with R0 (resection 0—microscopically clear margins) clearance. Operative outcomes to evaluate safety, which include either conversion to open procedure or change of planned surgery. This also included operation time and blood loss. Operative outcomes to evaluate safety, which include either conversion to open procedure or change of planned surgery. This also included operation time and blood loss. Postoperative outcomes, which looked at postoperative complications, length of hospital stay and length of chest drainage. Together, these domains formed a comprehensive framework for evaluating the clinical benefits of incorporating 3DCT into routine thoracic surgical planning. Postoperative outcomes, which looked at postoperative complications, length of hospital stay and length of chest drainage. Together, these domains formed a comprehensive framework for evaluating the clinical benefits of incorporating 3DCT into routine thoracic surgical planning. Results Our initial PubMed search identified 1,932 studies that mentioned 3DCT and after our initial screening, 121 were included based on the title and/or abstract. Further analysis resulted in the exclusion of seventy-three studies as seen in Figure 1. An additional eight were excluded as they were not related to anatomical lung resection. Of the 40 studies included in this review, 27 focused primarily on anatomical findings and variations as summarised in Table 2. Amongst these, 9 validated the 3DCT with intraoperative anatomy (8-11,13-15,29,33). Twenty-three studies directly evaluated the clinical impact of 3DCT reconstruction in preoperative planning and addressed a least one of the primary outcome measures. The remaining seventeen focused on anatomical description without any surgery performed, these studies were excluded from further analysis (12,17-28,30-32,34). Table 2Anatomical studiesYearStudyTypeNumber/powerIntraoperative comparisonAnatomy/lobe specificOutcome2008Fukuhara et al. (8)Retrospective49YesNoPA branches identification—95.2% identified2009Akiba et al. (9)Retrospective27YesNoPulmonary vessel and bronchi identification. PA branches—95% accuracy2014Hagiwara et al. (10)Retrospective179YesNoPulmonary vessel branching patterns PA branches identification—97.8%2016Yang et al. (11)Prospective10YesNoPA branches reconstructed up to Grade 52017Fourdrain et al. (12)Retrospective44NoNoPA branching patterns and variations bilaterally2017Le Moal et al. (13)Prospective9YesNoAnatomical accuracy with intraoperative comparison2018Smelt et al. (14)Prospective16YesNoNo statistical difference between 2DCT and 3DCT in determining PA branches2019Sardari Nia et al. (15)Prospective25YesNoAnatomical accuracy with intraoperative comparison2019Xu et al. (16)Retrospective133NoNoComparison study—but did not compare their primary endpoint (anatomical variation); 3DCT identified variation in 57.3% patients2020Isaka et al. (17)Retrospective103NoLULBranching patterns of intersegmental; PV compared to PA and bronchus—not independent of each other2021Liu et al. (18)Retrospective325NoNoISP identification using 3DCT with a cutting plane2021He et al. (19)Retrospective166NoLULAnatomical variation of lingula artery2021Wang et al. (20)Retrospective1,520NoUpper lobesDistributional features of bilateral superior PV2022Fan et al. (21)Retrospective136NoLULAnatomical consideration for LUL PA, PV and bronchus2022Zhong et al. (22)Retrospective212NoRULAnatomical consideration for RUL PA, PV and bronchus2022He et al. (23)Retrospective166NoLULBranching patterns of LUL bronchus2022Zhou et al. (24)Retrospective157NoRULAnalysis of intersegmental vein V2a to RUL2022Wang et al. (25)Retrospective179NoRight lungBranching patterns of PA2023Javed et al. (26)Retrospective10,000NoRMLBronchial branching to RML and any gender specific variations of the RML2023Ma et al. (27)Retrospective420NoBilateral lungsDistribution and variation of the PA system2024Liao et al. (28)Retrospective500NoLULDistribution pattern of V1+2 d in left superior PV2023Cannone et al. (29)Prospective11YesNoAnatomical accuracy with identification of bronchovascular structures2023Wei et al. (30)Retrospective608NoRMLRML venous anatomy and variations2024Xie et al. (31)Retrospective358NoNoPA branching pattern including a new classification system for training2024Zou et al. (32)Retrospective800NoRLLAnatomical variation of anatomy in S6 and S*2024Laven et al. (33)Prospective13YesNoAnatomical variations occurred in 62% of cases 2025Liu et al. (34)Retrospective77NoNoAccurate segmentecomy can be performed by identifying targeted segmental anatomy in preoperative planning2DCT, two-dimensional computed tomography; 3DCT, three-dimensional computed tomography; ISP, intersegmental plane; LUL, left upper lobe; PA, pulmonary artery; PV, pulmonary vein; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe; S*, segment*; S6, segment 6; V1+2, subsegmental vein branch V1+2; V2a, subsegmental vein branch V2a. Anatomical studies 2DCT, two-dimensional computed tomography; 3DCT, three-dimensional computed tomography; ISP, intersegmental plane; LUL, left upper lobe; PA, pulmonary artery; PV, pulmonary vein; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe; S*, segment*; S6, segment 6; V1+2, subsegmental vein branch V1+2; V2a, subsegmental vein branch V2a. Resection margin Nine studies met the inclusion criteria for reporting resection margins, as seen in Table 3 (6,15,29,35-40). Browne et al. reported change in surgical decision from segmentectomy to lobectomy due to 3DCT findings (10.5%) (6). Both Wang et al. and He et al. reported needing to increase their surgical margins by resecting an additional wedge when only using standard CT but not when using 3DCT (38,39). None of the trials reported missed nodules. There is a paucity of comparative evidence demonstrating superiority of 3DCT reconstruction in achieving adequate resection margins, but overall, the studies were small and heterogeneous with respect to segments removed and the location of the nodules. Table 3Resection marginsYearAuthorStudy designNumber/powerComparison 3DCT vs. 2DCTPrimary endpointOutcome2011Eguchi et al. (35)Retrospective analysis of 3DCT reconstruction in preoperative planning and intraoperative anatomical identification using an iPad14NoPositive resection margins and postoperative complicationsNo positive resection margins and perioperative or postoperative complications in all 14 cases2019Sardari Nia et al. (15)Prospective single-centre, single- arm study focused on preoperative planning25NoRadical resection, number of lymph nodes dissected and operation timeAll R0 resections with average margin 10 mm (5–33 mm); reconstructions revealed anatomic variations in 4 (15.4%) patients2021Liu et al. (36)Retrospective analysis to evaluate effect of inexperienced surgeons using 3DCT for surgical planning and investigating learning curve156NoPerioperative outcomes including operation time, blood loss, resection margin and postoperative complicationsAll R0 resections; learning curve is 30 cases for proficiency2022Lin et al. (37)Retrospective analysis of 3DCT reconstruction in nodule size and resection margin in segmentectomy18NoTumour size and its malignant correlation; 3DCT in preoperative planning by measuring resection marginCorrelation coefficient statistically significant for tumour size and malignancy; the correlation coefficient for measuring resection margin was statistically significant2022Wang et al. (38)Retrospective analysis comparing 3DCT analysis in completing complex segmentectomy97YesPerioperative outcomes including operation time, blood loss, postoperative complications2DCT arm included four patients requiring additional wedge for margin clearance compared to none required in the 3DCT arm; mean operation time was shorter in the 3DCT group (111.4±20.8 vs. 127.1±32.2 min), P=0.012023Cannone et al. (29)Prospective single-centre experience using 3DCT reconstruction in planning VATS segmentectomy11NoResection margin achieved with 3DCT reconstruction; accuracy in identifying bronchovascular structures with intraoperative comparisonOne case of lobectomy vs. segmentectomy due to margin involvement intraoperatively; anatomical accuracy with identification of bronchovascular structures2024He et al. (39)Retrospective analysis comparing perioperative outcomes and complications within 30 days of anatomical segmentectomy256YesPerioperative complications within 30 daysFive additional wedges in 2DCT arm—only one due to resection margins; reduced intraoperative blood loss and shorter postoperative hospital stay (P=0.005); higher number of LN sampling and more combined procedures (segmentectomy + sub-segmentectomy), P<0.001 in the 3DCT arm2024Bian et al. (40)Retrospective analysis using 3DCT reconstruction in preoperative planning to determine type of segmentectomy/sub-segmentectomy195NoPosition of nodules in relation to the anatomical resection requiredNodule relationship with adjacent intersegmental veins is what determines the specific surgical procedure, rather than their diameter and depth2024Browne et al. (6)A prospective single-centre cohort trial to determine safety and feasibility of using 3DCT reconstruction in robotic segmentectomy79NoRate of conversion of robotic segmentectomy to lobectomy after preoperative planning10.45% (7 patients) had change from segmentectomy to lobectomy and one intraoperatively because positive resection margin2DCT, two-dimensional computed tomography; 3DCT, three-dimensional computed tomography; LN, lymph node; R0, resection 0 (microscopically negative); VATS, video assisted thoracic surgery. Resection margins 2DCT, two-dimensional computed tomography; 3DCT, three-dimensional computed tomography; LN, lymph node; R0, resection 0 (microscopically negative); VATS, video assisted thoracic surgery. Operative outcomes Six studies reported conversion from video-assisted thoracic surgery (VATS) to open thoracotomy, and uniportal to multiport VATS when using 3DCT (6,9,10,33,41,42). A full breakdown is provided in Table 4. The reasons for conversion were reported as bleeding, adhesions and lymph node upstaging. Zhu et al., in a comparative study, reported 3 conversions (5.7%) in the 3DCT arm and 9 (6.2%) in the 2DCT arm. They performed propensity score matching and found no difference with the use of 3DCT (P=0.71) (45). Table 4Perioperative outcomesYearStudyConversion 3D armConversion 2D armOperation time (min)Blood loss (mL)Adverse events (Clavien-Dindo)Postoperative stay (days)Length of chest drainage (days)Prolonged air leak2008Fukuhara et al. (8)0Not applicableNot includedNot includedNot includedNot includedNot includedNot included2009Akiba et al. (9)2 due to N2 disease (reason unclear)Not applicableNot includedNot includedNot includedNot includedNot includedNot included2011Eguchi et al. (35)Procedure was already done via thoracotomy (cannot convert)Not applicable210±5321±15 Not includedNot includedNot includedNot included2014Hagiwara et al. (10)5 conversions to thoracotomy for bleeding (n=124)Not applicable230 [132–444]110 [0–1,406]10 (> Grade 2)Not includedNot included22014Chan et al. (43)Not includedNot applicableNot includedNot includedNot includedNot includedNot includedNot included2016Yang et al. (11)0Not applicable67.5±15.255.3±20.7Not included7.3±2.1 (standard practice patients to stay until wounds healed)Not includedNot included2017Yao et al. (44)1 due to calcified LN (n=63)Not applicable153 [95–250]112 [10–400]Not included6 [4–15]3 [1–7]22017Le Moal et al. (13)Not includedNot applicableNot includedNot includedNot includedNot includedNot includedNot included2018Smelt et al. (14)Not includedNot applicableNot includedNot includedNot includedNot includedNot includedNot included2019Sardari Nia et al. (15)0Not applicable281±56Not included11 > Grade 2Not includedNot included42019Xu et al. (16)Not includedNot applicable3D: 173.5±35.1 vs. 2D: 168.8±52.2, P=0.553D: 41.3±18.2 vs. 2D: 42.2±14, P=0.533D: 9 (> Grade 2) vs. 2D: 1 (> Grade 2), P=0.433D: 4.9±3.6 vs. 2D: 4.0±1.3, P=0.16Not included3D: 2; 2D: 12020Qiu et al. (41)003D: 116±30.7; 2D: 125.1±23.63D: 20.9±12.2 vs. 2D: 18.2±12.2, P=0.073D: 3; 2D: 13D: 4.7±1.5 vs. 2D: 5±1.5, P=0.213D: 4.1±1.5 vs. 2D: 4.1±1.4, P=0.753D: 1; 2D: 12021Zhu et al. (45)3, P>0.999, P>0.993D: 145.7±33.9 vs. 2D: 164.2±41.8, P=0.013D: 60.4±45.4 vs. 2D: 100.8±83.9, P=0.0093D: 10 (18.9%) vs. 2D: 27 (25.5%), P=0.333D: 4.24±1.84 vs. 2D: 4.59±2.41, P=0.343D: 2.75±1.92 vs. 2D: 2.97±2.16, P=0.523D: 4; 2D: 52021Liu et al. (36)0Not applicable119 [57–245]37 [15–247]11 (> Grade 2)3.7 [3–25]2.3 [1–23]52022Wang et al. (38)003D: 111.4±20.8 vs. 2D: 127.1±32.2, P=0.0073D: 47.9±29.1 vs. 2D: 51.1±36.3, P=0.643D: 11; 2D: 253D: 4.6±1.7 vs. 2D: 4.4±1.7, P=0.52Not included3D: 11.9% vs. 2D: 30.9%, P=0.032022Lin et al. (37)Not includedNot applicableNot includedNot includedNot included5 [5–7]Not includedNot included2023Cannone et al. (29)0Not applicable142 [105–182.5]Not includedNot included6 [5–7]4.4 (3-6) days12024Chen et al. (46)003D: 100 [85–120] vs. 2D: 100 [81–140], P=0.823D: 50 [20–100] vs. 2D: 50 [20–100], P=0.773D: 10 vs. 2D: 13, P=0.523D: 8 [7–8] vs. 2D: 7 [6–8], P=0.693D: 3 [2–3] vs. 2D: 3 [2–3], P=0.50Not included2024Laven et al. (33)1 (due to adhesions)Not applicableNot includedNot included5 (including 3 re-admissions)Not includedNot included12025Liu et al. (34)Not includedNot applicable180 [130–225]70 [50–100]3 (> Grade 2)7 [6–8]Not included92024He et al. (39)003D: 170.5±47.4 vs. 2D: 176±52.3, P=0.353D: 77.5±29 vs. 2D: 120±68.5, P<0.0013D: 12 vs. 2D: 19, P=0.0413D: 5.7±3.5 vs. 2D: 7.3±3.6, P=0.001Not included3D: 4, 2D: 32024Bian et al. (40)Not includedNot applicable160 [135–193]Not includedNot included4 [3.4–5.2]2.25 [2–3]182024Browne et al. (6)3 (2 for adhesions and 1 for LN)Not applicable17650 [25–100]19 (Grade 1), 21 (Grade 2), 2 (Grade 3), 1 mortality2 [1–5]2 [1–5]10 discharged with chest drainsData are shown as mean ± standard deviation or median [interquartile range] unless otherwise indicated. 2D, two-dimensional; 2DCT, two-dimensional computed tomography; 3D, three-dimensional; 3DCT, three-dimensional computed tomography; LN, lymph node; N2, level 2 lymph node. Perioperative outcomes Data are shown as mean ± standard deviation or median [interquartile range] unless otherwise indicated. 2D, two-dimensional; 2DCT, two-dimensional computed tomography; 3D, three-dimensional; 3DCT, three-dimensional computed tomography; LN, lymph node; N2, level 2 lymph node. Evaluating blood loss in isolation is challenging in a group of heterogeneous operations and without a control group. While the absolute values are presented in Table 4, this section focuses on studies that compared outcomes between the 3DCT reconstruction group and the 2DCT imaging group. Six comparative studies assessed blood loss across both groups (16,38,39,41,45,46). The only prospective randomised controlled trial, which included one hundred and ninety-one patients, reported no significant difference in intraoperative blood loss between the 3DCT and 2DCT arms (P=0.77) (46). However, Zhu et al. found that mean blood loss in the 3DCT group was 60.4±45.4 mL, compared to 100.8±83.9 mL in the 2DCT group (P=0.009) (45). Similarly, He et al., in a 2024 study involving 265 patients, observed a mean blood loss of 77.5±29 mL in the 3DCT group compared with 120±68.5 mL in the 2DCT group (P=0.04) (39). The remaining three comparative studies did not find a statistically significant difference in intraoperative blood loss between the two groups. The DRIVATS trial used operative time as its primary endpoint and reported no statistically significant difference between groups: median duration in the 3DCT arm was 100 minutes, which is the same as in the 2DCT arm (P=0.82) (46). However, among the five retrospective comparative studies, three reported a statistically significant reduction in operative time in the 3DCT reconstruction group compared to the 2DCT group, as detailed in Table 4 (39,41,45). However, the time difference between the two groups was only nine to eighteen minutes. Postoperative outcomes Common complications reported across studies included prolonged air leak, postoperative haemoptysis, pneumonia, haemothorax, empyema, and chylothorax. One single-armed study documented 32 out of 59 patients (54.2%) experienced a total of 42 adverse events, including a postoperative death due to multi-organ failure caused by thrombotic events. Most events were Clavien-Dindo Grade II or lower, with 4 events (8.3%) classified as Grade IIIa or higher (6). Two studies reported a single case of stroke in each (15,17). Among the comparative studies, the DRIVATS trial found no statistically significant difference in overall postoperative complications between the groups (46). One study demonstrated an overall reduction in postoperative complications in the 3DCT arm. They reported nineteen cases (16.2%) in the 2DCT group compared with 12 cases (8.1%) in the 3DCT group (P=0.04) (39). Hospital length of stay, readmissions, and need for further interventions served as additional postoperative markers. One centre in China stated that routine practice was to keep patients hospitalised until full wound healing was achieved, with an average stay of 7.3±2.1 days (11). Most other studies reported average stays ranging from 3 to 7 days. Browne et al. documented a median length of stay of 2 days (range, 1–5 days), although ten of their 79 patients were discharged with chest drains in situ (6). Laven et al. did not report length of stay but noted three readmissions (n=13), including two cases of subcutaneous emphysema resolving after 5 days and one empyema requiring re-intervention following a persistent air leak lasting more than 12 days (33). One comparative study, conducted by He et al., reported a statistically significant difference in hospital stay, with patients in the 3DCT group having a shorter average stay of 5.7±3.5 days compared to 7.3±3.6 days in the 2DCT group (P=0.001) (39). Data on chest tube duration was limited, with only eight studies reporting this outcome. Most indicated an average drainage duration of 2 to 3 days. Only three comparative studies reported on chest drainage duration, none of which demonstrated statistical significance (41,45,46). Comparative studies The findings from these studies are summarised in Table 5. The largest prospective, comparative study was the DRIVATS trial by Chen et al., which enrolled 191 patients (46). The two other prospective trials were smaller in scale: Smelt et al. reported 16 patients, and Laven et al. involved 13 patients (14,33). Laven et al. found 3DCT provided more accurate localisation of nodules at the segmental level in 5 out of 13 patients (38%). Furthermore, the operating surgeons indicated that they would have altered their surgical plans for eight patients based on 3DCT findings compared to standard 2DCT scans (33). Table 5Comparative studiesYearAuthorStudy designNumber/powerIntraoperativecorrelationOperation time (min)Blood loss (mL)Outcomes2DCT3DCTTotal2019Smelt et al. (14)Prospective analysis of the number of PA branches supplying lobe161616YesNot includedNot included3DCT scan had lower correlation when compared to 2DCT in analysing PA branching (P=0.143 vs. P=0.07), not statistically significant2019Xu et al. (16)Retrospective analysis of anatomical variations detected on 3DCT scans3796133No3D: 173.5±35.1 vs. 2D: 168±52.2, P=0.553D: 41.3±18.2 vs. 2D: 42.2±14, P=0.53A total of 73 structural variations were detected in 53 patients (57.3%), highest number in PA; no difference in operation time or postoperative outcomes2020Qiu et al. (41)Retrospective analysis of clinical outcomes in 3DCT vs. 2DCT and 3D printing in segmentectomy136131298No3D: 116±30.7; 2D: 125.1±23.63D: 20.9±12.2 vs. 2D: 18.2±12.2, P=0.07No statistical difference in operation time or intraoperative bleeding; 2 patients required an additional wedge resection in the 2DCT arm for inadequate resection margins; 4 patients had severe postoperative complications in the 3DCT arm vs. 1 in the 2DCT arm (not statistically significant)2021Zhu et al. (45)Retrospective propensity matched analysis comparing the use of 3DCT reconstruction in uVATS14553198Not well defined3D: 145.7±33.9 vs. 2D: 164.2±41.8, P=0.013D: 60.4±45.4 vs. 2D: 100.8±83.9, P=0.009Reduced operation time in the 3DCT arm (P=0.01); increased number of LN dissected in 3DCT arm (8.19±6.89 vs. 5.78±3.3, P=0.02); anatomical variations on 3DCT (not compared to 2DCT)—4 types arterial and 2 types venous amongst 8 patients (15%)2022Wang et al. (38)Retrospective study comparing 3DCT analysis in completing complex segmentectomy554297Some3D: 111.4±20.8 vs. 2D: 127.1±32.2, P=0.0073D: 47.9±29.1 vs. 2D: 51.1±36.3, P=0.64Mean operation time was shorter in the 3DCT group (P=0.01); 2DCT arm included 4 patients that required additional wedge resection for margin clearance compared to none in the 3DCT arm2024Chen et al. (46)Prospective multicentre RCT comparing clinical outcomes in segmentectomies using 3DCT vs. 2DCT9695191No3D: 100 [85–120] vs. 2D: 100 [81–140], P=0.823D: 50 [20–100] vs. 2D: 50 [20–100], P=0.77No difference in operation time in either arm (100 min, P=0.82); no difference in intraoperative blood loss (50 mL, P=0.77); no difference in postoperative hospital-stay or chest tube drainage2024Laven et al. (33)Prospective pilot study analysing preoperative planning of anatomical lung resection using 2DCT compared to 3DCT131313YesNot includedNot includedMore accurate information on location of nodule 5/13 (38%); surgeons would have altered their surgical plans for 8/13 (62%) patients based on 3DCT compared to 2DCT scans2024He et al. (39)Retrospective analysis comparing perioperative outcomes and complications within 30 days of anatomical segmentectomy117148256Not well defined3D: 170.5±47.4 vs. 2D: 176±52.3, P=0.363D: 77.5±29 vs. 2D: 120±68.5, P<0.001Reduced intraoperative blood loss and shorter postoperative hospital stay (P=0.005); higher number of LN sampling and more combined procedures (segmentectomy + sub-segmentectomy) (P<0.001); reported 5 additional wedges in the 2DCT arm—only 1 due to resection marginsData are shown as mean ± standard deviation or median [interquartile range] unless otherwise indicated. 2D, two-dimensional; 2DCT, two-dimensional computed tomography; 3D, three-dimensional; 3DCT, three-dimensional computed tomography; LN, lymph node; PA, pulmonary artery; RCT, randomised controlled trial; uVATS, uniportal video-assisted thoracic surgery. Comparative studies Data are shown as mean ± standard deviation or median [interquartile range] unless otherwise indicated. 2D, two-dimensional; 2DCT, two-dimensional computed tomography; 3D, three-dimensional; 3DCT, three-dimensional computed tomography; LN, lymph node; PA, pulmonary artery; RCT, randomised controlled trial; uVATS, uniportal video-assisted thoracic surgery. Of the remaining studies, three demonstrated clinically significant findings favouring the use of 3DCT. Zhu et al. conducted a retrospective, propensity-matched analysis performed via uniportal VATS. The study included 53 cases in the 3DCT group and 145 in the 2DCT group. The authors reported a significant reduction in operative time in the 3DCT arm, along with an increased number of lymph nodes resected (45). Similarly, He et al., in a 2024 retrospective study of 256 patients, reported a greater lymph node yield in the 3DCT arm, along with reduced intraoperative blood loss and a shorter length of hospital stay (39). A further study by Wang et al. also demonstrated a statistically significant reduction in operative time in the 3DCT group, as outlined in Table 5 (38). The authors did not allude to why they thought lymph node retrieval was better with the use of 3DCT. Discussion Numerous studies have explored the benefits of 3DCT reconstruction, yet few have demonstrated clear objective superiority over traditional 2DCT. Subjectively, most authors agree that 3DCT enhances understanding of tumour location and can identify anatomical variations pre-emptively. This improved spatial knowledge helps determine the actual operation and operative steps, reduces intraoperative uncertainty and potentially decreases operative time and the risk of bronchovascular injury. In studies with intraoperative anatomical comparison, concordance with the 3D plan is confirmed. The primary aim of pulmonary resections is to achieve not only R0 resection but also adequate margins. There was only one prospective single-armed study in which 10.5% (n=67) reported a change in surgical plan from segmentectomy to lobectomy to achieve adequate resection margin using 3DCT (6). This was further supported by Wang et al., reporting 4 patients (7%) requiring an additional wedge resection due to inadequate resection margins (38), and another reporting a single case of additional wedge being required (0.8%) in the 2DCT arm (39). In contrast, Cannone et al. reported the need to convert from segmentectomy to lobectomy intraoperatively in one patient despite the use of 3DCT (9%) (29). While there is some evidence supporting 3DCT usage to improve margin of resection, this is not conclusive. This is an important question that needs sufficiently powered, comparative studies. Conversion rates as a safety outcome were inconsistently reported, with only two comparative studies including this endpoint (6,45). One study found more conversions for bleeding in the 3DCT arm, which may reflect chance rather than causation (41). Similarly, while several studies reported lower intraoperative blood loss with 3DCT, the modest differences (e.g., 20–30 mL) is not clinically meaningful. Operation time is frequently cited as a proxy for efficiency, yet its value is context dependent. Factors such as the large variation in complexity between different segmentectomies, surgeon experience, and intraoperative events all play a part. While three retrospective comparative studies reported reduced operative time with 3DCT, averaging a 15–25 minutes saving, the prospective DRIVATS trial did not replicate this finding (38,41,45,46). Whether this time saving is clinically or economically meaningful remains open to interpretation. Postoperative outcomes, including postoperative complications, hospital stay or chest tube drainage, were well documented in just over half the studies, and were reported to be similar between those who use 3DCT and 2DCT scans. The most meaningful data comes from studies that directly compare 2DCT and 3DCT. Of the eight studies identified, five were retrospective and therefore subject to inherent bias. The only well-powered study demonstrating statistical significance was by Zhu et al., who found reduced operative time and increased lymph node retrieval in the 3DCT arm (45). However, the relationship between 3D imaging and nodal dissection has not been fully elucidated. In terms of level I evidence, there was only one randomised trial by Chen et al. It evaluated one hundred and ninety-one patients, using operative time as the primary endpoint. The study found no significant difference in operative duration, blood loss, postoperative complications, or hospital stay. The authors alluded that a higher proportion of simple segmentectomies in their cohort may have reduced the impact of 3DCT. While this explanation is plausible, it highlights the importance of stratification between disparate segmentectomy operations (46). This is consistent with there being no direct effect of 3DCT on clinical outcomes. The benefit of 3DCT may also extend beyond the operating theatre. In the multidisciplinary team setting, quick and accurate assessment of resectability is critical, particularly for borderline operable patients or those being considered for induction therapy. Misclassification of resectability can lead to treatment failure or missed opportunities for curative surgery. Rapid access to 3DCT reconstructions may help standardize surgical decision-making and reduce variability across centres. Finally, the Delphi consensus report published after the DRIVATS trial reflects expert opinion on best practices in minimally invasive segmentectomy. Among 21 panelists from three continents, there was strong agreement (94.4%) that 3DCT should be used for complex segmentectomies. The majority also supported its use intraoperatively and in preoperative nodule localisation. While consensus reports should be interpreted with caution and do not replace empirical data, they do reflect the evolving clinical consensus and practical realities of modern thoracic surgery (42). One of the principal limitations of this review was the paucity of high-quality comparative studies evaluating the clinical efficacy of 3DCT in anatomical lung resection. Robust comparative trials are essential to determine whether 3DCT truly confers an advantage over conventional CT scans. Most available studies are retrospective in design and thus susceptible to selection and reporting bias, whereas the few prospective, comparative studies that exist are often limited by small sample sizes and inadequate statistical power. Moreover, there is a notable lack of investigations assessing the clinically meaningful outcomes, such as resection margin status, detection of additional or missed nodules, and postoperative complications—parameters that are central to evaluating the value of anatomical lung resection. Finally, the heterogeneity of the surgical procedures analysed across studies, including segmentectomies, complex segmentectomies and lobectomies, further complicates the ability to draw definitive conclusions regarding the benefit of 3DCT in thoracic surgery. Looking ahead, further evidence is anticipated from the PATCHES trial, a prospective randomized controlled study that aims to compare 2DCT and 3DCT imaging in thoracic surgery, looking at negative margin (R0) resection rate, resection margin (staple line-to-tumor distance), and thoracotomy conversions. This may provide more definitive answers regarding the added value of 3DCT (47). Conclusions While 3DCT reconstruction offers clear subjective advantages in visualizing pulmonary anatomy and guiding surgical planning, current evidence does not conclusively demonstrate its superiority over conventional 2DCT in improving objective clinical outcomes. Retrospective studies suggest potential benefits such as enhanced spatial understanding, reduced intraoperative uncertainty, improved resection margins: however, these findings are limited by methodological heterogeneity, small sample sizes, and lack of standardized outcome measures. High-quality prospective data, including results from ongoing randomized controlled trials such as PATCHES study, are needed to determine whether 3DCT provides measurable improvements in surgical efficacy, safety and patient outcomes. Until such evidence becomes available, 3DCT should be regarded as a valuable adjunct, particularly for more complex procedures. 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. Footnotes Reporting Checklist: The authors have completed the Narrative Review reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1836/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-2025-1836/coif). Tim Batchelor has served on advisory boards for AstraZeneca, BMS and JnJ. He has also received honoraria from Medtronic, Intuitive, JnJ, Medela, BMS, AstraZeneca. The other authors have no conflicts of interest to declare. References Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1836/prf Supplementary Materials The article’s supplementary files as