Performance of multi-planar volume rendering versus maximum intensity projection and minimum intensity projection on pulmonary nodule detection

Introduction

Lung cancer is the leading cause of cancer-related death worldwide and in China, despite the increasing prevalence of lung cancer screening and advancements in treatment (1-3). Early and accurate detection of pulmonary nodules remains one of the most challenging aspects of managing early-stage lung cancer (4,5). Numerous studies have shown that early detection and timely diagnostic evaluation of pulmonary nodules can significantly reduce cancer-related mortality and improve the 5-year survival rate (6,7). However, missed or misinterpreted abnormalities in routine lung cancer screening using low-dose computed tomography (CT), as well as delays in follow-up diagnostic scans, have been shown to adversely affect the prognosis of early-stage lung cancer (8). Consequently, addressing the growing workload for radiologists caused by large volumes of thin-section CT images while maintaining diagnostic accuracy and efficiency is crucial.

Over the past decades, considerable efforts have been made to improve the diagnostic efficiency and accuracy of pulmonary nodules, including imaging techniques (9,10), computer-aided diagnosis (CAD) systems (11), diagnosis workflows (12,13), and post-processing techniques (14). Among these, post-processing techniques such as maximum intensity projection (MIP) and minimum intensity projection (MinIP) are widely available in most workstations and are seamlessly integrated into the diagnostic workflows of radiologists. Studies suggest that MIP significantly improves detection efficiency by enhancing the visualization of vessels in their longitudinal course, aiding in the differentiation between vessels and nodules (15-17). Our prior study demonstrated that MIP with a slab thickness of 10 mm enhances the diagnostic efficiency of solid nodules (SNs), while MinIP with a 3 mm slab thickness is more effective for detecting subsolid nodules (SSNs) (18). Despite these findings, there is limited evidence supporting the superiority of post-processing techniques for detecting part-solid nodules (PSNs).

Multiplanar volume rendering (MPVR) is an advanced three-dimensional (3D) volume reconstruction technique that enhances tissue differentiation by customizing the CT value-opacity curve, enabling the assignment of varying brightness and colors to different structures. In comparison with the traditional 3D reconstruction techniques, MPVR preserves depth information and spatial relationships while enabling precise discrimination and visualization of tissues with different densities through multiplanar reconstruction and a density threshold optimization algorithm. Previous study indicated the strong correlation of MPVR measurement with pathological findings on invasive component (19). Also, research evidence from a recent phantom study demonstrated the robust performance of MPVR in nodule detection even under low- or ultra-low dose settings (20). However, the diagnostic performance of MPVR in clinical practice remains largely unknown.

This study aims to evaluate the diagnostic benefits of MPVR and compare its performance with established techniques such as 10 mm MIP and 3 mm MinIP in detecting solid, part-solid, and pure ground-glass nodules (GGNs) across various size categories. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-102/rc).

Methods

Participants

This retrospective study reviewed a total of 990 chest CT images of patients with a prior diagnosis of pulmonary nodules between January 2023 and October 2023. The inclusion criteria were: complete preoperative chest CT with a thin slice (≤1 mm). The exclusion criteria were as follows: (I) patients with extensive pulmonary diseases (e.g., tuberculosis, pneumoconiosis, pneumonia, fibrosis, or edema) that could obscure nodule detection or mimic pulmonary nodules; (II) CT images with significant motion artifacts or poor quality that could hinder post-processing and accurate analysis; (III) patients with a history of malignant tumors (optional, commonly used); (IV) multifocal lesions that could complicate analysis (optional). The diagnostic criteria for pulmonary nodules were spherical or ellipsoidal structures ≤3 cm in size, with density higher than the surrounding parenchyma. We classified pulmonary nodules into three categories based on their density: SNs, PSNs, and GGNs.

After applying these criteria, a total of 183 patients [68 women and 115 men; mean age ± standard deviation (SD), 51±13 years; age range, 24–84 years] were included in the study. Ethics approval was obtained from the Institutional Review Board of The First Affiliated Hospital of Chongqing Medical University (No. K2024-057-01; Date: 2024/02/5), and the requirement for written informed consent for this retrospective study was waived. The study was conducted in accordance with the ethical standards of the responsible institution regarding human subjects, as well as in compliance with the Declaration of Helsinki and its subsequent amendments.

CT screening

All chest CT examinations were performed using a 128-MDCT scanner (Somatom Definition Flash; Siemens Healthcare, Erlangen, Germany) at the end of inspiration during a single breath-hold. All CT acquisitions were scanned from the lung apex to the base. Imaging parameters were as follows: 110–120 kVp tube voltage, 50–150 mAs tube current with automatic exposure control technology, 0.5 seconds rotation time, 5/5 mm image slice thickness and slice interval, 0.6×64 mm detector collimation, 1.0 pitch; and caudocranial scan direction. All images were reconstructed with a 1 mm slice thickness and 512×512 matrix, and a lung algorithm for the lung window image.

Image interpretation

Independent evaluations of different series of images were performed at a Picture Archiving and Communication System (PACS) workstation (Vue PACS, Carestream, version 12.2.6.3000020.) with standard lung window settings [width, 1,600 Hounsfield units (HU); level, –600 HU] for all post-processing groups, but the observers were allowed to moderately adjust the window setting to keep with their normal workflow. MPVR, MIP and MinIP were all integrated into the Carestream Vue PACS. The slab thickness of MIP and MinIP was set to 10 and 3 mm on the basis of our previous study (18), and slice gap was 0. As for MPVR, the CT values threshold of –800 HU was set to differentiate lung nodules and vessels from the lung background on the detection view of MPVR images (21-23), and the CT value threshold of –350 HU was set to differentiate the solid component and ground-glass background on the PSN view, which was also adapted from the literature (19).

This study employed a crossover design to evaluate the diagnostic performance of MPVR, MIP, and MinIP. Each patient’s CT images were reconstructed using all three techniques, and independent evaluations were conducted by two radiologists: one junior radiologist with 3 years of experience and one senior board-certified chest radiologist with more than 10 years of experience. Both observers had been trained to evaluate pulmonary nodules prior to the study and were familiar with the operation of the PACS workstation and the visual appearance of MPVR, MIP, and MinIP images. To minimize bias, image sets were presented randomly with respect to both the patient order and the order of the post-processing techniques (MPVR, MIP, or MinIP). Observers were required to annotate each detected nodule and compare it with the corresponding one on the 1 mm axial images to maximize diagnostic accuracy. They recorded the number, size, and location of nodules for each patient. A minimum interval of two weeks between analyses of different sets of images for the same patient was implemented to mitigate memory effects.

Reference standard

The reference standard was established as the maximum number of nodules detected by the two readers, which were then further confirmed by a consensus group using 1 mm axial images. The consensus group, blinded to the results of the post-processing techniques, evaluated the nodules across multiple planes (axial, coronal, and sagittal) following established guidelines (24,25). Each detected nodule was classified as either a true- or false-positive. Discrepancies in nodule evaluation were resolved through joint discussion within the consensus group.

Overall, the study identified 1,018 true-positive SNs, 146 true-positive PSNs, and 583 true-positive GGNs in 183 patients. True-positive nodules were further categorized by size, with 1,266 nodules (72.5%) measuring <5 mm, 434 (24.8%) measuring 5–10 mm, and 47 (2.7%) measuring >10 mm. A total of 133 nodules (7.1%) were classified as false-positives by the consensus group. The false-positive rates for observer 1 and observer 2 were as follows: MPVR (106 and 93), MIP (95 and 78), and MinIP (124 and 101). Notably, within each post-processing group, the maximum number of false-positives per patient for either observer did not exceed two, and there were no statistically significant differences among the techniques.

Statistical analysis

Statistical analysis was performed using SPSS 22.0 (IBM Corp., Armonk, NY, USA) and JASP [JASP Team (2020), Version 0.14.1]. Statistical analysis included Wilcoxon matched-pairs signed-rank tests to compare detection rates across techniques for nodules of different sizes (<5 mm, 5–10 mm, >10 mm) and densities (solid, part-solid, and GGNs). Bayesian Wilcoxon signed-rank tests quantified the relative evidence [Bayes factor (BF₁₀)] for the alternative hypothesis; BF₁₀ >3 was taken as moderate evidence of a clinically relevant effect, complementing the frequentist P value. This approach was adopted because BF₁₀ directly quantifies evidence for the alternative or null hypothesis, which are not less sensitive to sample size. Interobserver agreement was assessed using kappa statistics, with κ>0.8 considered excellent. Bonferroni correction was applied to control for multiple comparisons, and a P value of <0.05 and BF₁₀ >3 were considered statistically significant.

Results

Interobserver agreement

The interobserver agreement for MPVR was predominantly excellent across all categories, with κ=0.913 [95% confidence interval (CI): 0.878–0.948] for the total nodules, κ=0.733–0.899 for density subgroups, and κ=0.651–0.909 for size subgroups (Table S1). Similarly, MIP demonstrated high interobserver agreement, with κ=0.949 (95% CI: 0.930–0.968) for the total nodules, κ=0.865–0.948 for density subgroups, and κ=0.668–0.892 for size subgroups. MinIP also achieved comparable results, with κ=0.972 (95% CI: 0.962–0.982) for the total nodules, κ=0.790–0.900 for density subgroups, and κ=0.700–0.912 for size subgroups. These results indicate a consistently high level of agreement among observers across all post-processing techniques.

Overall nodule detection performance

MPVR demonstrated the highest overall sensitivity (observer 1: 82%; observer 2: 86%), outperforming MIP (observer 1: 64%; observer 2: 69%) and MinIP (observer 1: 61%; observer 2: 64%) (Table 1, Table S2). MPVR detected significantly more nodules in both observer evaluations (P<0.001, BF₁₀ >1,000), reinforcing its superior diagnostic capability across all nodule categories (Table 2).

Nodule detection in size, location, and density subgroups

MPVR demonstrated consistently high sensitivity across all nodule sizes, with particularly strong performance for small nodules (<5 mm), achieving 84–90% sensitivity compared to 66–71% for MIP and 58–61% for MinIP (P<0.05, BF₁₀ >3). For medium nodules (5–10 mm), MPVR remained superior to MIP and comparable to MinIP, while all modalities performed near 100% for large nodules (>10 mm) (Tables 2,3, Table S3). By anatomical location, MPVR outperformed both alternative techniques across all regions, with significant advantages over MinIP in the left upper lobe (P<0.05, BF₁₀ >3) (Tables 1,2) Analysis by density revealed MPVR’s superior detection of SNs (89–94%) and PSNs (82–87%), whereas MinIP was more effective for GGNs, especially those 5–10 mm in size (91–99% vs. 37–41% for MPVR and 43–46% for MIP; P<0.05, BF₁₀ >3). For small GGNs (<5 mm), MPVR and MinIP showed comparable performance (Tables 2,3, Table S3).

Discussion

This study demonstrates that both observers detected significantly more pulmonary nodules using MPVR, highlighting its potential to improve the overall diagnostic performance compared to the well-established 10 mm MIP and 3 mm MinIP techniques on thin-section chest MDCT scans. MPVR was particularly effective in detecting nodules smaller than 5 mm, a subgroup where its sensitivity was notably higher than that of the other techniques. Although differences in detection rates did not reach statistical significance in some size subgroups, MPVR showed a trend toward higher detection sensitivity for PSNs compared to MIP and MinIP. These findings position MPVR not merely as an alternative post-processing tool but as a comprehensive visualization platform capable of supporting nuanced diagnostic tasks in routine CT interpretation. Notably, MinIP exhibited better sensitivity for GGNs, while MPVR performed comparably in detecting small GGNs (<5 mm). The excellent interobserver agreement observed with MPVR further supports its reproducibility and reliability in clinical practice.

Given the widespread use of MDCT for lung cancer screening, the number of incidentally detected pulmonary nodules has significantly increased (26). While MDCT is known to improve the detection rates of pulmonary nodules by offering higher spatial resolution and reducing partial-volume effects, it also introduces perception errors due to the large volume of thin-section data (27). Post-processing techniques such as MIP and MinIP have been widely studied for their ability to enhance diagnostic sensitivity and efficiency (17,18,28). However, these techniques exhibit certain limitations in clinical practice: MIP primarily improves the detection of SNs, whereas MinIP is more effective for SSNs. Both techniques have questionable capabilities in detecting nodules with varying densities, which limits their broader applicability (18,29). Previous evidence suggests that MIP is not sufficiently sensitive for nodules with diverse densities, with an overall sensitivity of only 0.60–0.70 for both junior and senior radiologists, findings consistent with those of our study (15,16). Furthermore, there is limited research on the overall detection efficiency of MinIP. Our study addresses this gap, showing that MinIP achieved detection rates of 0.49–0.52, 0.45–0.50, and 0.84–0.89 for SNs, PSNs, and GGNs, respectively. In contrast, MPVR demonstrated consistently high detection efficiency across all density subgroups. This advantage can be attributed to optimized thresholds for structure opacity and pseudo-color, which enhance the realistic visualization of pulmonary nodules. As outlined in our prior phantom and real-world studies (19), a CT value threshold of −800 HU was used in the detection view of MPVR images to differentiate nodules and vessels from the lung background. This approach simultaneously displayed solid and SSNs while minimizing background interference (Figure 1). Additionally, a dedicated PSN view with a CT value threshold of −350 HU was employed to define the solid components of PSNs, which were highlighted in red to simulate vessel density, consistent with previous findings (19) (Figure 1). Our results further suggest that MPVR detected significantly more nodules in the left upper lobe compared to MinIP. This difference likely reflects MPVR’s superior performance in identifying SNs, particularly small ones, which MinIP often misses in this region. However, the influence of nodule localization on the detection performance of post-processing techniques remains controversial (28,30). Therefore, these findings should be interpreted with caution and warrant further investigation.

Figure 1 Thin-slab 1 mm axial CT view (left), nodule detection view of MPVR (middle), and PSN view of MPVR (right) from the CT data set of a 60-year-old patient with a part-solid nodule (white boxes). In the PSN view of the MPVR images, the solid component within the part-solid nodule is displayed in red, consistent with the color of pulmonary vessels. Recent findings indicate that measuring the solid component on the PSN views of MPVR achieves high accuracy and consistency compared to pathological size, underscoring its clinical utility for nodule characterization. CT, computed tomography; MPVR, multi-planar volume rendering; PSN, part-solid nodule.

As a well-established post-processing technique, MPVR is typically pre-integrated into most diagnostic workstations. Previous phantom studies have demonstrated its excellent nodule detection performance, diagnostic robustness and efficiency under low-dose and ultra-low-dose scanning conditions. Our current study further indicates that, compared to MIP and MinIP, MPVR offers comprehensive nodule detection capabilities, including significantly improved sensitivity for small nodules and potentially enhanced detection of PSNs, while maintaining excellent interreader agreement comparable to that of MIP and MinIP. These advantages suggest that MPVR could serve as a primary screening tool for an initial comprehensive review, given its high overall sensitivity across nodule densities and sizes. Subsequently, MinIP images may be specifically consulted when pure GGNs are suspected on MPVR or when there is high clinical suspicion for GGNs. Furthermore, as a 3D post-processing technique, MPVR preserves depth information and accurately depicts spatial relationships between lesions and surrounding tissues, whereas MIP and MinIP, as flattened projections, are susceptible to structural overlap. Prior studies have confirmed that MPVR provides reliable lesion measurement capabilities, unlike MIP and MinIP, which are primarily limited to nodule detection (20). By offering both comprehensive detection and accurate measurement, MPVR has the potential to enhance radiologists’ reading efficiency and reduce overall interpretation time in lung cancer screening programs. The MPVR technique demonstrated significantly better performance in detecting small pulmonary nodules measuring less than 5 mm (Figure 2). Current guidelines and consensus statements on pulmonary nodule management emphasize the initial size of the nodule, noting that larger nodules have a higher probability of malignancy, while the probability for smaller nodules is relatively low (12,31). However, management protocols vary in terms of size thresholds and follow-up strategies, which complicates the accurate measurement of nodule growth over time (13,24,25). Despite the low malignancy rate of <5 mm nodules, their accurate detection remains clinically significant. Studies have shown that delayed follow-up of small nodules can adversely impact long-term survival, with reports indicating a 12.4% decrease in lung cancer cure rates for nodules initially measuring 4 mm when follow-up was delayed by 1 year (32). Similarly, prospective studies analyzing large databases revealed a substantial weekly mortality increase of 3.2% associated with delays in early lung cancer treatment. Importantly, the impact of delayed follow-up was found to be more pronounced in early-stage lung cancer compared to late-stage disease (33). These findings underscore the potential benefits of MPVR in improving the prognosis of early-stage lung cancer. By enabling the inclusion of most nodule candidates with varying sizes and densities in a single enhancement tool, MPVR facilitates more comprehensive nodule detection. Additionally, its application simplifies radiologists’ workflows and reduces workload, further supporting its utility in routine clinical practice.

Figure 2 Thin-slab 1 mm axial CT view (upper left), MPVR view (upper right), MIP view (bottom left), and MinIP view (bottom right) from the CT data set of a 63-year-old patient with a small solid nodule (white boxes). MPVR images visualize the lung background and the nodule using distinct opacity and pseudo-color settings, preserving spatial allocation while reducing the influence of “anatomic noise”. CT, computed tomography; MinIP, minimum intensity projection; MIP, maximum intensity projection; MPVR, multi-planar volume rendering.

Currently, researchers are increasingly focusing on the use of artificial intelligence (AI)-based detection tools, also known as computer-aided detection (CADe), to enhance the diagnostic efficiency of pulmonary nodules and reduce the rates of missed or misdiagnosed nodules (34). While these emerging CADe systems have demonstrated significant benefits and high efficiency in nodule detection (35-37), their integration into routine clinical workflows remains challenging compared to the established use of post-processing techniques (27,38). In clinical practice, radiologists typically begin by quickly reviewing 5 mm axial images and post-processed images to identify potential nodule candidates, followed by a detailed evaluation of thinner-sliced images for further characterization (30). However, CADe systems are often not integrated into PACS software, requiring radiologists to switch to external platforms and repeat the image reading process. Moreover, CADe tools are unable to present nodules in the order that aligns with individual radiologist preferences, which may increase the workload and disrupt workflow efficiency during nodule diagnosis. Interestingly, several studies have developed CADe or deep learning (DL) models based on post-processed images, demonstrating superior performance compared to raw CT images (38-40). Inspired by these findings and the results of our study, future research combining different post-processing techniques (MPVR, MIP, MinIP) with AI is expected to further enhance the diagnostic performance of pulmonary nodules. Such integration could leverage the complementary strengths of both approaches, potentially streamlining workflows while improving accuracy and efficiency in nodule detection.

Limitation

There are several limitations in this study. Firstly, the case-enriched design created a homogenous cohort by systematically excluding complex pulmonary pathologies (e.g., tuberculosis, pneumoconiosis, pneumonia, fibrosis, or edema) and poor-quality images. While this enhanced internal validity, it also introduced a potential selection bias, thereby may reduce the study’s external validity and real-world clinical relevance. Secondly, only two observers were included, and reading times were unrestricted, though the excellent interobserver agreement suggests consistent results. Future studies should involve more observers and assess reading speed across techniques. Thirdly, the single-center retrospective design restricts external validity. Further multi-center studies with prospective design are needed to validate these findings across diverse clinical settings. Lastly, future research could explore the integration of pot-processing techniques like MPVR with AI-based detection tools to enhance diagnostic accuracy and workflow efficiency in lung cancer screening.

Conclusions

In conclusion, MPVR can significantly improve the overall detection rates of pulmonary nodules in comparison with the well-established 10 mm MIP and 3 mm MinIP, especially for nodules smaller than 5 mm. In addition, MPVR showed the potential to facilitate the detection of PSNs and had satisfying inter-observer agreement. Therefore, MPVR is a useful and promising diagnostic tool for thoracic MDCT to highlight most nodules with different sizes and densities.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-102/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-102/dss

Funding: This work was supported by the National Key R&D Program of China (grant No. 2020YFA0714002), Chongqing Health Appropriate Technology Promotion Project (No. 2023jstg044), the Key Project of Technological Innovation and Application Development of Chongqing Science and Technology Bureau (No. CSTC2021jscx-ksbN0030), and the Joint Project of Chongqing Health Commission and Science and Technology Bureau (No. 2022ZDXM006).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-102/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of The First Affiliated Hospital of Chongqing Medical University (No. K2024-057-01. Date: 2024/02/5) and the requirement for written informed consent for this retrospective analysis was waived.

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

References

1. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, Bray F. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021;71:209-49. [Crossref] [PubMed]
2. Xia C, Dong X, Li H, Cao M, Sun D, He S, Yang F, Yan X, Zhang S, Li N, Chen W. Cancer statistics in China and United States, 2022: profiles, trends, and determinants. Chin Med J (Engl) 2022;135:584-90. [Crossref] [PubMed]
3. Li N, Tan F, Chen W, Dai M, Wang F, Shen S, et al. One-off low-dose CT for lung cancer screening in China: a multicentre, population-based, prospective cohort study. Lancet Respir Med 2022;10:378-91. [Crossref] [PubMed]
4. Liang M, Tang W, Xu DM, Jirapatnakul AC, Reeves AP, Henschke CI, Yankelevitz D, Low-Dose CT. Screening for Lung Cancer: Computer-aided Detection of Missed Lung Cancers. Radiology 2016;281:279-88. [Crossref] [PubMed]
5. Aberle DR, Adams AM, Berg CD, Black WC, Clapp JD, Fagerstrom RM, Gareen IF, Gatsonis C, Marcus PM, Sicks JD. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med 2011;365:395-409. [Crossref] [PubMed]
6. Blandin Knight S, Crosbie PA, Balata H, Chudziak J, Hussell T, Dive C. Progress and prospects of early detection in lung cancer. Open Biol 2017;7:170070. [Crossref] [PubMed]
7. de Koning HJ, van der Aalst CM, de Jong PA, Scholten ET, Nackaerts K, Heuvelmans MA, et al. Reduced Lung-Cancer Mortality with Volume CT Screening in a Randomized Trial. N Engl J Med 2020;382:503-13. [Crossref] [PubMed]
8. Zuniga PVS, Ost DE. Impact of Delays in Lung Cancer Treatment on Survival. Chest 2021;160:1934-58. [Crossref] [PubMed]
9. Lucero MY, Chan J. Photoacoustic imaging of elevated glutathione in models of lung cancer for companion diagnostic applications. Nat Chem 2021;13:1248-56. [Crossref] [PubMed]
10. Liu M, Li J, Zhang D, Yuan K, Wu F, Yang F, Peng G, Tang B, Orlandini LC. Challenges of Treating Lung Cancer Patients at MR-Linac Using MR-Based Synthetic CT Calculation in the Adaptive Workflow. Int J Radiat Oncol 2023;117:e684.
11. Jenkin Suji R, Bhadauria SS, Wilfred Godfrey W. A survey and taxonomy of 2.5D approaches for lung segmentation and nodule detection in CT images. Comput Biol Med 2023;165:107437. [Crossref] [PubMed]
12. Mazzone PJ, Lam L. Evaluating the Patient With a Pulmonary Nodule: A Review. JAMA 2022;327:264-73. [Crossref] [PubMed]
13. Eguchi T, Adusumilli PS. Risk stratification for lung nodules: Size isn’t everything. J Thorac Cardiovasc Surg 2017;153:1557-62. [Crossref] [PubMed]
14. Walsh SL, Nair A, Hansell DM. Post-processing applications in thoracic computed tomography. Clin Radiol 2013;68:433-48. [Crossref] [PubMed]
15. Park EA, Goo JM, Lee JW, Kang CH, Lee HJ, Lee CH, Park CM, Lee HY, Im JG. Efficacy of computer-aided detection system and thin-slab maximum intensity projection technique in the detection of pulmonary nodules in patients with resected metastases. Invest Radiol 2009;44:105-13. [Crossref] [PubMed]
16. Peloschek P, Sailer J, Weber M, Herold CJ, Prokop M, Schaefer-Prokop C. Pulmonary nodules: sensitivity of maximum intensity projection versus that of volume rendering of 3D multidetector CT data. Radiology 2007;243:561-9. [Crossref] [PubMed]
17. Gruden JF, Ouanounou S, Tigges S, Norris SD, Klausner TS. Incremental benefit of maximum-intensity-projection images on observer detection of small pulmonary nodules revealed by multidetector CT. AJR Am J Roentgenol 2002;179:149-57. [Crossref] [PubMed]
18. Li WJ, Chu ZG, Zhang Y, Li Q, Zheng YN, Lv FJ. Effect of Slab Thickness on the Detection of Pulmonary Nodules by Use of CT Maximum and Minimum Intensity Projection. AJR Am J Roentgenol 2019;213:562-7. [Crossref] [PubMed]
19. Li WJ, Chu ZG, Li D, Jing WW, Shi QL, Lv FJ. Accuracy of solid portion size measured on multiplanar volume rendering images for assessing invasiveness in lung adenocarcinoma manifesting as subsolid nodules. Quant Imaging Med Surg 2024;14:1971-84. [Crossref] [PubMed]
20. Chen S, Gao L, Tan M, Zhang K, Lv F. Performance Evaluation of Four Deep Learning-Based CAD Systems and Manual Reading for Pulmonary Nodules Detection, Volume Measurement, and Lung-RADS Classification Under Varying Radiation Doses and Reconstruction Methods. Diagnostics (Basel) 2025;15:1623. [Crossref] [PubMed]
21. Shi Z, Deng J, She Y, Zhang L, Ren Y, Sun W, Su H, Dai C, Jiang G, Sun X, Xie D, Chen C. Quantitative features can predict further growth of persistent pure ground-glass nodule. Quant Imaging Med Surg 2019;9:283-91. [Crossref] [PubMed]
22. Bak SH, Lee HY, Kim JH, Um SW, Kwon OJ, Han J, Kim HK, Kim J, Lee KS, Quantitative CT. Scanning Analysis of Pure Ground-Glass Opacity Nodules Predicts Further CT Scanning Change. Chest 2016;149:180-91. [Crossref] [PubMed]
23. Eguchi T, Kondo R, Kawakami S, Matsushita M, Takeda T, Sakaizawa T, Tominaga Y, Saito G, Hashizume M, Toishi M, Shiina T, Yoshida K, Amano J, Koizumi T. Pulmonary pure ground-glass opacity lesions: Use of computed tomography attenuation for predicting tumor progression. J Clin Oncol 2013;31:7548.
24. Bankier AA, MacMahon H, Goo JM, Rubin GD, Schaefer-Prokop CM, Naidich DP. Recommendations for Measuring Pulmonary Nodules at CT: A Statement from the Fleischner Society. Radiology 2017;285:584-600. [Crossref] [PubMed]
25. Chen B, Li Q, Hao Q, Tan J, Yan L, Zhu Y, et al. Malignancy risk stratification for solitary pulmonary nodule: A clinical practice guideline. J Evid Based Med 2022;15:142-51. [Crossref] [PubMed]
26. Kim TJ, Kim CH, Lee HY, Chung MJ, Shin SH, Lee KJ, Lee KS. Management of incidental pulmonary nodules: current strategies and future perspectives. Expert Rev Respir Med 2020;14:173-94. [Crossref] [PubMed]
27. Svoboda E. Artificial intelligence is improving the detection of lung cancer. Nature 2020;587:S20-2. [Crossref] [PubMed]
28. Kawel N, Seifert B, Luetolf M, Boehm T. Effect of slab thickness on the CT detection of pulmonary nodules: use of sliding thin-slab maximum intensity projection and volume rendering. AJR Am J Roentgenol 2009;192:1324-9. [Crossref] [PubMed]
29. Azour L, Ko JP, Naidich DP, Moore WH. Shades of Gray: Subsolid Nodule Considerations and Management. Chest 2021;159:2072-89. [Crossref] [PubMed]
30. Kilburn-Toppin F, Arthurs OJ, Tasker AD, Set PA. Detection of pulmonary nodules at paediatric CT: maximum intensity projections and axial source images are complementary. Pediatr Radiol 2013;43:820-6. [Crossref] [PubMed]
31. Gould MK, Tang T, Liu IL, Lee J, Zheng C, Danforth KN, Kosco AE, Di Fiore JL, Suh DE. Recent Trends in the Identification of Incidental Pulmonary Nodules. Am J Respir Crit Care Med 2015;192:1208-14. [Crossref] [PubMed]
32. Yankelevitz DF, Yip R, Henschke CI. Impact of Duration of Diagnostic Workup on Prognosis for Early Lung Cancer. J Thorac Oncol 2023;18:527-37. [Crossref] [PubMed]
33. Khorana AA, Tullio K, Elson P, Pennell NA, Grobmyer SR, Kalady MF, Raymond D, Abraham J, Klein EA, Walsh RM, Monteleone EE, Wei W, Hobbs B, Bolwell BJ. Time to initial cancer treatment in the United States and association with survival over time: An observational study. PLoS One 2019;14:e0213209. [Crossref] [PubMed]
34. Bi WL, Hosny A, Schabath MB, Giger ML, Birkbak NJ, Mehrtash A, Allison T, Arnaout O, Abbosh C, Dunn IF, Mak RH, Tamimi RM, Tempany CM, Swanton C, Hoffmann U, Schwartz LH, Gillies RJ, Huang RY, Aerts HJWL. Artificial intelligence in cancer imaging: Clinical challenges and applications. CA Cancer J Clin 2019;69:127-57. [Crossref] [PubMed]
35. Aggarwal R, Sounderajah V, Martin G, Ting DSW, Karthikesalingam A, King D, Ashrafian H, Darzi A. Diagnostic accuracy of deep learning in medical imaging: a systematic review and meta-analysis. NPJ Digit Med 2021;4:65. [Crossref] [PubMed]
36. Yang L, Cai H, Luo X, Wu J, Tang R, Chen Y, Li W. A lightweight neural network for lung nodule detection based on improved ghost module. Quant Imaging Med Surg 2023;13:4205-21. [Crossref] [PubMed]
37. Jirapatnakul A, Yip R, Myers KJ, Cai S, Henschke CI, Yankelevitz D. Assessing the impact of nodule features and software algorithm on pulmonary nodule measurement uncertainty for nodules sized 20 mm or less. Quant Imaging Med Surg 2024;14:5057-71. [Crossref] [PubMed]
38. Zheng S, Guo J, Cui X, Veldhuis RNJ, Oudkerk M, van Ooijen PMA. Automatic Pulmonary Nodule Detection in CT Scans Using Convolutional Neural Networks Based on Maximum Intensity Projection. IEEE Trans Med Imaging 2020;39:797-805. [Crossref] [PubMed]
39. Wu Y, Qi Q, Qi S, Yang L, Wang H, Yu H, Li J, Wang G, Zhang P, Liang Z, Chen R. Classification of COVID-19 from community-acquired pneumonia: Boosting the performance with capsule network and maximum intensity projection image of CT scans. Comput Biol Med 2023;154:106567. [Crossref] [PubMed]
40. Gil J, Choi H, Paeng JC, Cheon GJ, Kang KW. Deep Learning-Based Feature Extraction from Whole-Body PET/CT Employing Maximum Intensity Projection Images: Preliminary Results of Lung Cancer Data. Nucl Med Mol Imaging 2023;57:216-22. [Crossref] [PubMed]