Value of CT diagnostic techniques based on imaging post-processing systems in the early diagnosis and treatment of lung cancer

Abstract

Objective: To evaluate the application value of CT diagnostic technology based on the Shukun Imaging Post-Processing System for early screening and diagnosis of lung cancer. Methods: A total of 35 patients diagnosed with lung cancer postoperatively and 53 patients with benign nodules were included in this retrospective study, all of whom were treated in the Department of Thoracic and Cardiovascular Surgery of the Second Affiliated Hospital of Fujian Medical University from January 2020 to December 2023. All patients underwent chest spiral CT examinations. Original thin-slice axial CT images were processed using Shukun software for three-dimensional reconstruction of the lesions, surrounding lung tissue, and trachea. The diagnoses and malignant risk indicators of pulmonary nodules were established based on the final imaging results. Results: Statistical analysis showed that the sensitivity of Shukun processing technology in diagnosing early-stage lung cancer was 82.86%, with a specificity of 88.46%, when compared to postoperative pathological analysis. Univariate logistic regression analysis indicated that features such as burr sign, lobulation sign, pleural traction sign, vascular convergence sign, vacuole sign, and nodule size derived from Shukun processing had significant predictive value for malignant nodules (P<0.05). Conclusion: Shukun processing technology can effectively reconstruct ordinary CT tomographic images into three-dimensional representations, enhancing the visualization of spatial relationships between the tumor and adjacent anatomical structures, including trachea, pleura, bronchi, and blood vessels. It has high clinical diagnostic value for the early diagnosis of malignant pulmonary nodules.

Introduction

A pulmonary nodule is defined as a solid lung lesion measuring less than 3 cm in diameter [1,2]. These small nodules are often incidentally identified through chest X-rays or high-resolution CT scans. Their characteristics can vary, including benign, malignant, or atypical hyperplasia, which represents an intermediate state between the two [3,4]. Benign pulmonary nodules may originate from infections, inflammations, or other non-neoplastic conditions, while malignant nodules typically indicate precancerous or early-stage lung cancer, with treatment and prognosis closely linked to the timely detection. The relationship between pulmonary nodules and lung cancer is intricate; not all pulmonary nodules progress to lung cancer [5,6].

Statistical evidence indicates that pulmonary nodules smaller than 5 mm have a very low probability of developing into lung cancer, whereas the risk of malignancy increases with nodule size [7,8]. Currently, the primary imaging modality for lung cancer diagnosis is spiral CT. However, conventional axial CT images often fail to provide accurate qualitative assessments of lung tissue lesions, resulting in a low rate of early lung cancer diagnosis. Shukun Imaging Post-Processing software enhances standard CT scans by reconstructing images into three-dimensional representations, thereby clarifying the anatomy of pulmonary nodules, surrounding tissues, and blood vessels, which may improve the detection rate of malignant tumors [9,10]. Based on this background, this study aims to evaluate the diagnostic efficacy of this technology in 88 patients with pulmonary nodules, contributing to enhanced approaches for diagnosing malignant lung tumors.

Materials and methods

Patient selection

Between January 2022 and June 2023, a total of 35 patients with lung cancer and 53 patients with benign nodules were admitted to the Department of Thoracic and Cardiovascular Surgery at the Second Affiliated Hospital of Fujian Medical University.

This retrospective study aimed to estimate the sample size for two groups in a parallel 1:1 design, dividing patients into a malignant group and a benign group. Based on previous comparisons of diagnostic effects using similar image post-processing systems for pulmonary nodules, the average diagnostic coincidence rates were μ1 = 0.8 for the control group and μ2 = 0.9 for the observation group, with a standard deviation of s = 0.1. Accounting for a 10% dropout rate and setting the type I error rate (α) at 0.05 with a confidence level of (1-β) = 80%, the required sample size was calculated using the formula:

$$\mathrm{n}1=\mathrm{n}2=\frac{2{(\mathrm{Z}\mathrm{a}/2+\mathrm{Z}\mathrm{\beta })}^2\times {\mathrm{\sigma }}^2}{{(\mathrm{\mu }1-\mathrm{\mu }2)}^2}$$

Substituting the values (μ1 = 0.8, μ2 = 0.9, σ = 1, α = 0.05, β = 0.2), the calculations yielded n1 = n2 = 12, leading to a minimum sample size of 24. To enhance the reliability of the results, all eligible patients during the study period, totaling 88 cases, were included.

Inclusion criteria

1. Preoperative chest CT indicating nodules with a diameter greater than 8 mm; 2. Presence of common imaging features of nodules, including spiculation, calcification, satellite lesions, pleural traction sign, lobulation sign, and vascular sign; 3. Availability of complete standard medical records, including current and past history, as well as comprehensive preoperative laboratory and imaging examination results; 4. Completion of chest and related examinations at our hospital.

Exclusion criteria

1. Patients with concurrent liver, lung, or gastrointestinal tumors identified during preoperative examinations; 2. Patients exhibiting multiple nodules in both lungs on chest CT, where pathology cannot be determined; 3. Patients with well-defined pulmonary masses suspected to be advanced tumors with associated thrombosis or compression; 4. Patients experiencing significant organ dysfunction, such as heart or kidney failure; 5. Patients with mental illness or speech disorders that impede normal communication; 6. Patients with incomplete clinical data.

Data acquisition

Data were obtained by reviewing the electronic medical record system, including patient blood test results and CT scan images. The chest CT examinations were conducted using a 64-slice spiral CT scanner with the following parameters: tube voltage of 120 kV, tube current of 300 mA, pitch of 1.375:1, scan layer thickness of 5.0 mm, matrix of 512×512, and reconstruction layer thickness of 1.25 mm. A comprehensive lung scan was performed for all patients based on the above parameters. After the examination, the images were uploaded to a central workstation, where the Shukun Imaging Post-Processing System performed three-dimensional reconstructions. Imaging diagnostic specialists then reviewed the reconstructed images to identify relevant pulmonary tissues. The system automatically generated parameters related to the lesions, including size, shape, and location. A preliminary diagnostic assessment for the lesion was subsequently formulated based on these imaging characteristics.

Observation indicators

Primary observation indicators: The final diagnostic criteria for lung cancer, which included pathological examination of lung lesions, evaluation of tissue morphology, and staining and immunohistochemical analysis to determine the pathological classification of lung nodules, were employed to compare the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the automatic diagnostic results generated by the Shukun processing software, thereby evaluating its diagnostic efficacy. Additionally, the predictive capacity of various imaging characteristics of pulmonary nodules for distinguishing between benign and malignant outcomes was assessed [11].

Secondary observation indicators: Comparisons were made of peripheral blood test results between the benign and malignant groups, including routine blood tests, liver and kidney function assessments, and lung cancer markers, to explore the efficacy of hematological indicators in diagnosing malignant lung tumors.

Data analysis

Data analysis was performed using SPSS 22.0 software. Measurement data were expressed as mean ± SEM. The t-test was used for analyzing measurement data with a normal distribution. Paired sample t-tests were employed for within-group comparisons, while independent sample t-tests were applied for between-group comparisons. For data that did not follow a normal distribution, the Mann-Whitney U test was used. Count data were expressed as rates (percentages) and analyzed using the x ² test. Variables with statistically significant differences in univariate analysis were included in logistic regression analysis to calculate the odds ratio (OR) and its 95% confidence interval (CI). A P-value of less than 0.05 was considered statistically significant.

Results

Final pathological reports of the 88 cases

Among the 88 cases examined, 35 patients (39.78%) were pathologically diagnosed with lung cancer, while 53 patients (60.22%) were diagnosed with non-lung cancer conditions. The types and composition of lung cancer cases are shown in Figure 1.

Figure 1.

Pathological classification of lung cancer patients.

Automatic diagnostic results of pulmonary nodules using Shukun processing software

The analysis of the CT/Shukun system processing technology revealed a sensitivity of 82.86%, a specificity of 88.46%, a positive predictive value of 80.56%, and a negative predictive value of 86.80% for early-stage lung cancer diagnosis, indicating a relatively high diagnostic value for this technology. The area under the curve (AUC) was 0.848, with a 95% CI of 0.758 to 0.938 (Table 1 and Figure 2). Typical images processed using the CT/Shukun system are shown in Figure 3.

Table 1.

The diagnostic efficiency of lung nodule Shukun processing software

		CT/Shukun System		Total
		Malignant	Benign
Pathological examination results	Malignant	29	6	35
	Benign	7	46	53
	Total	36	52	88

Figure 2.

The AUC diagram of the diagnosis of pulmonary nodules by Shukun system.

Figure 3.

Classic image of lung nodules. Three-dimensional reconstruction of pulmonary nodules using CT plain scan and Shukun post-processing technology. Panels (A) and (D) show the horizontal view of the left lower pulmonary nodule, panels (B) and (E) display the coronal view, and panels (C) and (F) present the three-dimensional reconstruction of the lung nodule generated by the post-processing system.

Comparison of baseline data between malignant and benign groups

No statistically significant differences were found in gender, age, and BMI between the two groups, indicating comparability (Table 2).

Table 2.

Comparison of baseline data between the two groups

Group	Gender (Male/Female)	Age	BMI	Hypertension	Nodule Distribution (Left/Right)	Time of Nodule Detection
Malignant group (n = 35)	20/15	37.8±5.2	23.2±2.7	7	18/17	2.8±0.4
Benign group (n = 53)	30/23	38.0±5.9	24.1±2.8	5	29/24	2.5±0.5
Statistical value	0.002	-0.163	-1.497	1.202	0.092	-0.992
P	0.960	0.871	0.138	0.273	0.762	0.324

Comparison of serological data between malignant and benign groups

Similarly, the comparison of serological data indicated no statistically significant differences in gender, age, and BMI between the two groups, indicating comparability (Table 3).

Table 3.

Comparison of serum related examinations between the two groups

Group	WBC (×109)	AST (U/L)	ALT (U/L)	Scr (mmol/L)	NEU (×109)	LYM (×109)
Malignant group (n = 35)	9.88±1.2	20.5±2.4	22.6±2.7	71.6±4.8	5.45±2.25	1.39±0.62
Benign group (n = 53)	9.65±1.4	22.2±2.8	21.9±2.9	72.0±5.0	5.37±2.16	1.32±0.43
Statistical value	0.797	-1.213	1.139	-0.373	0.167	0.626
P	0.428	0.228	0.258	0.710	0.868	0.533

Comparison of serum tumor marker levels between malignant and benign groups

Analysis of serum tumor markers revealed no statistically significant differences in peripheral blood levels of carcinoembryonic antigen (CEA), neuron-specific enolase (NSE), and cytokeratin fragment 19 (CYFRA21-1) (P = 0.773, 0.666 and 0.720, respectively) (Figure 4).

Figure 4.

Comparison of serum tumor markers CEA, NSE, and CYFRA21-1 between the two groups. Note: CEA: carcinoembryonic antigen, NSE: neuron-specific enolase; CYFRA21-1: cytokeratin fragment 19.

Differences in imaging features between malignant and benign groups

Statistically significant differences were observed between the malignant and benign groups in terms of spiculation, lobulation, pleural traction sign, vascular convergence sign, vacuole sign, size, calcification, and satellite lesions. Multivariate analysis identified these factors as independent influencing factors for the malignancy of pulmonary nodules (Tables 4, 5 and Figures 5, 6).

Table 4.

Univariate analysis of malignant transformation in patients with pulmonary nodules

CT Findings	Benign group (n = 53)	Malignant group (n = 35)	F/χ2	P
Spiculation	13	19	8.066	0.005
Lobulation	16	19	5.110	0.024
Pleural traction sign	11	17	7.518	0.006
Vascular convergence sign	12	16	5.173	0.023
Vacuole sign	8	15	8.416	0.004
Nodule size
>1 cm	44	20	7.396	0.006
≤1 cm	9	15
Calcification	16	22	9.170	0.002
Satellite lesions	10	18	10.302	0.001

Table 5.

Logistic analysis of pulmonary nodule malignancy

Variable	Standardized β	OR	95% CI	P
Spiculation	0.884	3.46	1.47-8.89	0.038
Lobulation	0.747	3.21	1.36-7.26	0.032
Pleural traction sign	0.558	4.27	1.15-7.48	0.020
Vascular convergence sign	0.487	4.42	1.26-8.15	0.002
Vacuole sign	0.665	3.73	1.17-9.13	0.024
Nodule size >1 cm	0.849	3.71	1.05-8.66	0.005
Calcification	-3.023	0.49	0.32-0.64	0.047
Satellite lesions	-4.328	0.85	0.44-0.77	0.004

Figure 5.

Satellite focus and calcification of pulmonary nodules. A: Pulmonary nodules satellite focus; B: Calcification image.

Figure 6.

Images of pulmonary nodular lobulation sign, pleural traction sign, vascular convergence sign and vacuole sign. A: Vacuole sign; B: Lobulation sign; C: Vascular convergence sign; D: Pleural traction sign.

Discussion

Pulmonary nodules generally refer to small, round or oval lesions within the lung parenchyma, typically measuring less than 3 cm in diameter. When a pulmonary nodule exceeds this threshold, it is classified as a pulmonary mass [12]. The nature of these nodules can be benign or malignant; benign nodules may arise from inflammation, infection, or benign tumors, whereas malignant nodules are usually associated with lung cancer [13-15]. Therefore, the early diagnosis of the nature of pulmonary nodules is crucial for improving patient prognosis. Epidemiological studies show that the detection rate of pulmonary nodules has significantly increased with the widespread use of high-resolution CT (HRCT). Among populations undergoing chest CT scans, the proportion of detected pulmonary nodules can exceed 50% [16-18].

Since its introduction in the 1970s, CT scanning technology has undergone revolutionary advancements in the medical field. Compared to traditional X-ray examinations, CT provides more detailed and clearer images of internal structures, particularly in complex organs like the lungs. By utilizing a high-speed rotating X-ray emitter and receiver, CT scans can capture images from different angles, which are then processed by computer algorithms to reconstruct three-dimensional, well-defined images. This technology allows doctors to observe the lung’s overall structure and accurately locate, measure, and analyze the characteristics of pulmonary nodules, making it the preferred method for lung examinations [19,20]. However, conventional axial CT images often fail to provide effective qualitative diagnoses of lung tissue lesions, leading to lower rates of early lung cancer detection [21,22]. In contrast, the Shukun Imaging Post-Processing System enhances conventional CT scans by reconstructing and processing images to present pulmonary nodules and surrounding tissues and blood vessels in a three-dimensional anatomical format. This advancement improves the detection rate of malignant lung tumors.

Previous studies have demonstrated that image post-processing systems can significantly enhance the diagnostic accuracy of malignant pulmonary nodules. The Shukun system employs innovative three-dimensional reconstruction technology based on standard spiral CT scans, distinguishing it from other CT post-processing techniques such as MPR, MIP, and CRP. Myrian software, for example, plays a critical role in clinical screening, diagnosis, disease assessment, and surgical guidance. By offering three-dimensional reconstructions, it provides a more intuitive and comprehensive view of the anatomical structures of lesions, adjacent tissues, and blood vessels. This capability not only increases the early detection rate of tumors but also offers valuable guidance for surgical intervention [23]. In our study, the Shukun Imaging Post-Processing System was used to diagnose 88 patients with pulmonary nodules. Pathological confirmation revealed that 35 patients had lung cancer, while 55 had benign lesions. Statistical analysis showed that the Shukun Post-Processing System achieved a sensitivity of 82.86%, specificity of 88.46%, positive predictive value of 80.56%, and negative predictive value of 86.80%. The AUC was 0.848, with a 95% CI of 0.758 to 0.938, indicating that the Shukun Imaging Post-Processing System possesses good diagnostic value for lung cancer. These findings are consistent with results reported in previous studies [23,24].

In addition, previous imaging studies have demonstrated that various imaging features can aid in differentiating between benign and malignant tumors, including the burr sign, lobulation sign, pleural traction sign, vascular convergence sign, and vacuole sign, which are commonly associated with malignancies. Conversely, calcification is typically indicative of benign lesions [25]. Our analysis of imaging data revealed statistically significant differences between benign and malignant pulmonary nodules in terms of spiculation, lobulation, pleural traction sign, vascular convergence sign, vacuole sign, size, calcification, and satellite lesions. Logistic regression analysis indicated that these characteristics serve as independent imaging factors for the malignancy of pulmonary nodules, whereas calcification and satellite lesions were identified as protective factors. The underlying mechanisms for these imaging characteristics are as follows: Spiculation occurs due to tumor cell infiltration into adjacent bronchovascular sheaths or local lymphatics, or as a result of the stimulation of connective tissue formation by tumor-associated fibrous bands. Radiographically, this manifests as short, radiating, unbranched lines extending from the edge of the lung mass or nodule into the surrounding parenchyma, without connecting to the pleura. Lobulation is primarily due to the uniform density of peripheral lung cancer. The pleural traction sign occurs when nodules exert a pulling effect on the pleura, while the vascular convergence sign is mainly characterized by thickened vascular structures, reflecting the increased blood supply needed for rapidly growing nodules. The vacuole sign manifests as 1-3 mm low-density areas within the nodule due to residual air-containing lung tissue or bronchi. The size of the pulmonary nodule correlates directly with the likelihood of malignancy. Studies have reported that nodules with a maximum diameter of 1 cm are more likely to be malignant, as malignant nodules generally exhibit more rapid growth than benign ones. Our study also showed that calcification and satellite lesions are protective factors against the malignancy of pulmonary nodules. Calcification in pulmonary nodules is commonly seen in post-inflammatory nodules, such as those resulting from tuberculosis or pulmonary hamartomas, which are often benign. Satellite lesions, smaller nodules located near the primary nodule, resemble satellites around the primary nodule and are typically indicative of benign conditions. These findings are consistent with the conclusions of previous researchers [25,26].

Previous studies have demonstrated that serological markers in patients diagnosed with early-stage lung cancer often exhibit no significant abnormalities [27]. In our study, we further explored the correlation between serological markers and malignant lung tumors. Our results showed no statistically significant differences in peripheral blood routine tests and tumor markers, including CEA, NSE, and CYFRA21-1, between the two groups. This finding is consistent with previous conclusions regarding the limited diagnostic efficacy of serological markers for lung cancer, aligning with current clinical practices where serum markers are not widely used for early screening of lung cancer [27,28]. However, some studies still suggest that tumor markers, such as CEA, NSE, and CYFRA21-1, possess significant predictive value, likely associated with larger tumor sizes and more advanced stages of the patients in these studies. Therefore, effective serum tumor markers could potentially play an important role in the diagnosis, treatment, and prognosis of lung cancer.

In conclusion, the Shukun Imaging Post-Processing System, through three-dimensional reconstruction of standard CT images of the lungs, demonstrates good diagnostic value for lung cancer. However, this study has the following limitations: certain findings require further validation through larger sample sizes in subsequent studies, and improving the efficacy of serological markers in lung cancer diagnosis remains a crucial area for future research to strengthen the conclusions of this research.

Disclosure of conflict of interest

None.