Study on the Diagnostic Value of Clinical Decision Rules for Malignant Tumors Complicated with Pulmonary Thromboembolism: A Retrospective Analysis

Abstract

Background

Recently released clinical decision rules (CDR) have demonstrated strong performance in diagnosing pulmonary thromboembolism (PTE) within the general population; however, there is a notable lack of comparative analysis regarding the diagnosis of PTE in patients with malignant tumor.

Objective

To evaluate the diagnostic efficacy of CDR for PTE in patients with malignant and non-malignant tumors.

Methods

A retrospective analysis was conducted on 537 patients who underwent CT pulmonary arteriography (CTPA) for suspected PTE at Yantaishan Hospital from October 2014 to March 2024. Clinical characteristics of the patients were collected, clinical decision rule scores were calculated, and their diagnostic value was assessed.

Results

In the malignant tumor group, the area under the curve (AUC) for the Wells score, revised Geneva score, age-adjusted d-dimer score (AADD), Years score, pulmonary embolism graduated d-dimer score (PEGeD), and the 4-level clinical pretest probability score (4PEPS) were 0.644, 0.584, 0.662, 0.663, 0.671, and 0.665, respectively, all below 0.7. The diagnostic performance of the Wells, AADD, Years, PEGeD, and 4PEPS score was superior to that of the revised Geneva score. In the non-malignant tumor group, the AUC values for the Wells score, revised Geneva score, AADD, YEARS score, PEGeD score, and 4PEPS score were 0.711, 0.699, 0.748, 0.756, 0.762, and 0.750, respectively. All AUC values exceeded 0.7 except for the revised Geneva score.

Conclusion

CDR was less effective in diagnosing PTE in patients with malignant tumors and was more suitable for those with non-malignant tumors.

Keywords

clinical decision rules malignant tumor pulmonary thromboembolism

Introduction

Pulmonary thromboembolism (PTE) was a clinical condition characterized by the obstruction of the pulmonary arteries and their branches due to blood clots originating from the right heart or the venous system. PTE and deep vein thrombosis (DVT) were collectively referred to as venous thromboembolism (VTE). In western countries, the incidence of VTE was as high as 6.25%, with an average of approximately 2335 individuals dying from VTE each day.¹ PTE was the leading cause of sudden death among patients with VTE, accounting for 10% of in-hospital fatalities.¹ CT pulmonary angiogram (CTPA) is currently the most widely used method for diagnosing PTE. However, this test was expensive and posed risks related to radiation exposure and allergic reactions.² Additionally, it was not suitable for certain patient populations, including pregnant women and individuals with renal failure. Some scholars had developed clinical decision rules (CDR) for the early diagnosis of PTE to reduce the need for CTPA examinations by comprehensively analyzing the clinical characteristics of PTE patients. Among these, the Wells score³ and the revised Geneva score⁴ were the most commonly used.

Patients with malignant tumors had a nine-fold higher risk of VTE and a two- to three-fold higher mortality rate compared to the general population.⁵ Early recognition of pulmonary embolism significantly improved survival rates in this patient group. The Wells score and the revised Geneva score, while highly effective in diagnosing pulmonary embolism in the general population,⁶ were less suitable for patients with malignant tumors.⁷ Researchers had developed several new scoring criteria, including the age-adjusted d-dimer (AADD) score,⁸ Years score,⁹ pulmonary embolism graduated d-dimer (PEGeD) score,¹⁰ and the 4-level clinical pretest probability score (4PEPS).¹¹ Compared to the Wells score and the revised Geneva score, the AADD, Years, and PEGeD scores demonstrated higher diagnostic accuracy or specificity in patients with non-malignant tumors.^12,13 However, there is a notable lack of comparative studies examining the efficacy of these new scores in diagnosing PTE in patients with malignant tumors.

The aim of this study was to evaluate the diagnostic value of the Wells score, revised Geneva score, AADD score, Years score, PEGeD score, and 4PEPS for the diagnosis of PTE in patients with malignant tumors, and to compare these findings with CDR for the diagnosis of PTE in patients with non-malignant tumors.

Methods

Study Protocol

This study was a retrospective cohort study conducted in accordance with the STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) guidelines.¹⁴ Data were obtained from the electronic medical records of 537 patients who were admitted to Yantaishan Hospital for suspected PTE and underwent CTPA during the inclusion period from October 2014 to March 2024.

Inclusion criteria: (1) patients who have completed CPTA examination; (2) patients who have completed D-dimer and oxygen saturation tests.

Exclusion criteria: (1) patients with incomplete medical records; (2) patients who had previously experienced PTE prior to the diagnosis of a malignant tumor; (3) patients who had experienced malignant tumor embolism, air embolism, or amniotic fluid embolism; and (4) patients who had received routine anticoagulant therapy (such as low-molecular-weight heparin, standard heparin, rivaroxaban, or warfarin) before admission.

Diagnostic criteria for tumors: Observe for signs of malignant tumors using computed tomography (CT), magnetic resonance imaging (MRI), or ultrasound examinations. Then, diagnose whether the tumor was malignant or benign through histological or cytological analysis.

The diagnostic criteria for PTE were outlined in the “Guidelines for the Diagnosis, Treatment, and Prevention of Pulmonary Embolism (2018)”. We gathered factors potentially associated with PTE through insights from clinical practice and an analysis of existing literature. Demographic and clinical data, including age, gender, medical history, and laboratory test results, were collected from electronic medical records. This study protocol was approved by the Ethics Committee of Yantaishan Hospital (Ethics Number: 2025079).

Patient Stratification

The 537 patients were divided into two groups: a malignant tumor group and a non-malignant tumor group, based on the presence or absence of malignant tumor indicators identified through imaging examinations such as CT, MRI, and ultrasound, as well as histological and cytological analyses. Subsequently, these two groups were further subdivided into PTE and non-PTE subgroups based on CTPA examination results. Clinical characteristics of the patients were collected, and statistical differences between the groups were analyzed based on the CDR scoring criteria.

Clinical Decision Rules Scoring Criteria

The criteria for the CDR scoring were presented in Table 1.

Table 1.

Clinical Decision Rules Scoring Criteria.

Scores Name	Items	Clinical Decision Rule Points	Decision Rule
Wells score¹⁵	previous PTE or DVT	1.5	Sum ≤ 4 and D-dimer < 0.5 mg/L: PTE excluded； Sum > 4: PTE not excluded
	heart rate >100 beats/min	1.5
	surgery or immobilization within the past 4 weeks	1.5
	hemoptysis	1
	active malignant tumor	1
	clinical signs of DVT	3
	alterative diagnosis less likely than PTE	3
the revised Geneva score⁴	age older than 65 years	1	Sum ≤ 5 and D-dimer < 0.5 mg/L: PTE excluded； Sum > 5: PTE not excluded
	previous deep venous thrombosis or pulmonary embolism	3
	surgery or fracture within 1 month	2
	active malignant tumor	2
	unilateral lower limb pain	3
	hemoptysis	2
	heart rate of 75 to 94 beats/min	3
	95 beats/min or more	5
	pain on lower-limb deep venous palpation and unilateral edema	4
AADD score¹⁶	NA	NA	50 years or younger: Wells score ≤ 4 and D-dimer ≤ 0.5 mg/L or > 50 years: D-dimer < age × 0.01 mg/L: PTE excluded
Years score¹⁷	clinical signs of DVT	1	0 score and D-dimer < 1 mg/L: PTE excluded; ≥ 1 score and D-dimer < 0.5 mg/L: PTE excluded
	hemoptysis	1
	alterative diagnosis less likely than PTE	1
PEGeD score¹⁸	NA	NA	Sum ≤ 4 and D-dimer < 1 mg/L: PTE excluded; Sum > 4 and ≤6 and D-dimer < 0.5 mg/L: PTE excluded; Sum > 6: PTE not excluded
4PEPS score¹⁹	age < 50 years	−2	Sum < 0 score: PTE excluded; Sum > 0 and ≤5 and D-dimer < 1 mg/L: PTE excluded; Sum > 6 and ≤12 and D-dimer < age × 0.01 mg/L: PTE excluded; Sum > 12: PTE not excluded
	50 ≤ age ≤ 64 years	−1
	chronic respiratory diseases	−1
	heart rate < 80 beats/min	−1
	chest pain and acute respiratory distress	1
	male	2
	receiving estrogen therapy	2
	previous deep venous thrombosis or pulmonary embolism	2
	syncope	2
	surgery or immobilization within the past 4 weeks	2
	blood oxygen saturation < 95%	3
	unilateral lower limb pain or edema	3
	alterative diagnosis less likely than PTE	5

Quality Control

The results of the CTPA were diagnosed with the assistance of an imaging physician and subsequently reviewed by the physician's supervisor. The scoring item “alternative diagnosis less likely than PTE” was evaluated by two physicians independently. In cases of inconsistent results, the supervising physician of one of the evaluators reassessed the findings.

Data Analysis

In the clinical characteristics, measurement data that follow a normal distribution were expressed as the mean ± standard deviation, and the t-test was used to compare differences between two groups. Data that do not follow a normal distribution were expressed as the median (25th, 75th percentiles), and the Mann-Whitney U test was used to compare differences between two groups. Enumeration data data were expressed as frequencies (%). If the sample size was >40 cases and the frequency was >5, the chi-square test (χ²) was used to compare differences between groups. If the sample size was >40 cases but the frequency was <5, the corrected chi-square test was applied. If the sample size was ≤40 cases, or the frequency was <1, Fisher's exact test was used to compare differences between groups. Additionally, when comparing clinical data across different tumor types, one-way ANOVA was used for multiple group comparisons of measurement data that followed a normal distribution, with the Least Significant Difference (LSD) test applied for two-group comparisons. For measurement data that did not follow a normal distribution, the Kruskal-Wallis test was employed for multiple group comparisons, and for enumeration data, one-way ANOVA was used for multiple group comparisons. Clinical predictive scores for malignant and non-benign tumor groups were statistically analyzed, with calculations of AUC values, sensitivity, and specificity. ROC curves were plotted, and the AUC were compared. Statistical analyses were conducted using SPSS version 25.0, and ROC curve analysis was performed with GraphPad Prism version 10.1.2. A confidence interval (CI) of 95% was used, and P <0.05 was considered statistically significant.

Results

Clinical Characteristics Analysis

There were 124 individuals in the malignant tumor group, of whom 64 were diagnosed with PTE and were included in the PTE subgroups, while the remaining 60 were placed in the non-PTE subgroups. Statistical analysis of the results revealed significant differences between the two subgroups regarding pain on “lower-limb deep venous palpation and unilateral edema”, “surgery or immobilization within the past 4 weeks”, and “D-dimer levels”. A total of 413 individuals were included in the non-malignant tumor group, of whom 142 were diagnosed with PTE and subsequently enrolled in the PTE subgroups. The remaining 271 individuals were placed in the non-PTE subgroups. A significant difference was observed between the two subgroups regarding “heart rate”, “lower-limb deep venous palpation and unilateral edema”, and “D-dimer levels” (P < 0.05), as illustrated in Table 2. In addition, statistically significant differences were observed in “heart rate”, “hemoptysis”, and “D-dimer levels” between the malignant and non-malignant tumor groups (P < 0.05), as presented in Table 3.

Table 2.

Comparative Analysis of Clinical Features of PTE and non-PTE Between the Two Groups.

	Malignant Group (124)			Non-malignant Tumor Group (413)
	PTE (64)	Non-PTE (60)	P value	PTE (142)	Non-PTE (271)	P value
Male	34 (53.1)	31(51.7)	0.871	78 (54.9)	142 (52.4)	0.624
Age	68 ± 10	67 ± 11	0.803	68 ± 12	66 ± 14	0.240
Heart rate (beats/min)	91 ± 17	88 ± 20	0.457	85 ± 17	81 ± 15	0.029
Chronic respiratory diseases	14 (21.9)	17 (28.3)	0.407	23 (16.2)	58 (21.4)	0.206
Previous PTE or DVT	7 (10.9)	6 (10.0)	0.865	14 (9.9)	14 (5.2)	0.072
Surgery or immobilization within the past 4 weeks	17 (26.6)	5 (8.3)	0.008	23 (16.2)	27 (10.0)	0.065
Chest pain	11 (17.2)	9 (15.0)	0.741	30 (21.1)	67 (24.7)	0.413
Acute respiratory distress	48 (75.0)	38 (63.3)	0.159	99 (69.7)	201 (74.2)	0.335
Hemoptysis	9 (14.1)	7 (11.7)	0.691	10 (7.0)	15 (5.5)	0.542
Syncope	5 (7.8)	2 (3.3)	0.490	7 (4.9)	11 (4.1)	0.681
Unilateral lower limb pain or edema	15 (23.4)	6 (10.0)	0.046	36 (25.4)	9 (3.3)	<0.001
D-dimer (mg/L)	2.69 (1.07, 4.64)	0.74 (0.33, 2.16)	<0.001	2.03 (0.94, 4.79)	0.44 (0.18, 1.18)	<0.001
Oxygen saturation < 95%	34 (53.1)	39 (65.0)	0.179	83 (58.5)	165 (60.9)	0.631

Note: Age and heart rate were expressed as mean ± standard deviation; D-dimer was expressed as the median (25th, 75th percentiles); other indicators were expressed as frequency (%).

Table 3.

Comparison of Clinical Characteristics Between the Malignant Tumor and Non-malignant Tumor Group.

	Malignant Group (124)	Non-malignant Tumor Group (413)	P value
Male	65 (52.4)	220 (53.3)	0.868
Age	67 ± 11	66 ± 14	0.446
Heart rate (beats/min)	89 ± 19	83 ± 16	<0.001
Chronic respiratory diseases	31 (25.0)	81 (19.6)	0.195
Previous PTE or DVT	13 (10.5)	28 (6.8)	0.173
Surgery or immobilization within the past 4 weeks	22 (17.7)	50 (12.1)	0.106
Chest pain	20 (16.1)	97 (23.5)	0.082
Acute respiratory distress	86 (69.4)	300 (72.6)	0.476
Hemoptysis	16 (12.9)	25 (6.1)	0.012
Syncope	7 (5.6)	18 (4.4)	0.551
Unilateral lower limb pain or edema	21 (16.9)	45 (10.9)	0.072
D-dimer (mg/L)	1.45 (0.59, 3.60)	0.80 (0.26, 2.08)	<0.001
Oxygen saturation < 95%	73 (58.9)	248 (60.0)	0.815

Note: Age and heart rate were expressed as mean ± standard deviation; D-dimer was expressed as the median (25th, 75th percentiles); other indicators were expressed as frequency (%).

Comparison of the Diagnostic Value of CDR for Malignant Tumors Combined with PTE Versus non-Malignant Tumors Combined with PTE

In the malignancy group, the revised Geneva score demonstrated the highest sensitivity for diagnosing PTE at 96.88%. The 4PEPS exhibited the highest specificity at 51.67% and the highest positive predictive value (PPV) at 64.20%. Additionally, the revised Geneva score had the highest negative predictive value (NPV) at 85.71%. However, all CDRs had an area under the curve (AUC) of less than 0.7. The most effective combined prediction was the PEGeD score, which achieved an AUC of 0.671, as illustrated in Table 4 and Figure 1A. In the non-malignant tumor group, the revised Geneva score exhibited the highest sensitivity at 92.96%, while the 4PEPS demonstrated the highest specificity at 69.00% and the highest positive predictive value at 57.79%. Additionally, the revised Geneva score had the highest negative predictive value at 92.70%. The AUC of all scores was greater than 0.7 except for the revised Geneva score. The PEGeD score yielded the best combined diagnostic result, with an AUC of 0.762, as illustrated in Table 4 and Figure 1B.

Figure 1.

ROC Curve of CDR in Malignant Group (A) and non-Malignant Tumor Group (B).

Table 4.

The Diagnostic Value of CDR in Malignant or non-Malignant Tumor Complicated with PTE.

Group	CDR	AUC	95% CI	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
Malignant tumor	Wells score	0.644	(0.545, 0.742)	93.75	35.00	60.61	84.00
	Revised Geneva score	0.584	(0.483, 0.685)	96.88	20.00	56.36	85.71
	AADD score	0.662	(0.565, 0.759)	89.06	43.33	62.64	78.79
	Years score	0.663	(0.566, 0.760)	85.94	46.67	63.22	75.68
	PEGeD score	0.671	(0.575, 0.768)	85.94	48.33	63.95	76.32
	4PEPS score	0.665	(0.568, 0.761)	81.25	51.67	64.20	72.09
Non-malignant tumor	Wells score	0.711	(0.661, 0.761)	90.14	52.03	49.61	90.97
	Revised Geneva score	0.699	(0.649, 0.749)	92.96	46.86	47.83	92.70
	AADD score	0.748	(0.700, 0.796)	88.03	61.62	54.59	90.76
	Years score	0.756	(0.708, 0.805)	84.51	66.79	57.14	89.16
	PEGeD score	0.762	(0.714, 0.810)	85.21	67.16	57.62	89.66
	4PEPS score	0.750	(0.700, 0.799)	80.99	69.00	57.79	87.38

Patients in the malignant tumor group exhibited the following distribution of affected systems: respiratory system (54/124), digestive system (25/124), reproductive system (19/124), breast (10/124), urinary system (7/124), endocrine system (2/124), hematological system (2/124), nervous system (1/124), and other systems (4/124). Due to the limited sample size of tumors in other systems, the scoring results were not significant. Therefore, we conducted subgroup analyses exclusively for tumors of the respiratory, digestive, and reproductive systems (Table 5). The results indicated that the AADD score exhibited the highest predictive value for respiratory system tumors (AUC = 0.692). In contrast, for digestive system tumors, the Wells score, AADD score, and 4PEPS score demonstrated the best predictive value (AUC = 0.549). For reproductive system tumors, the Years score and PEGeD score performed the best, with an AUC of 0.767.

Table 5.

The Diagnostic Value of CDR in Malignant Tumor in Different Sites Complicated with PTE.

Tumor	CDR	AUC	95% CI	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
Respiratory system tumor	Wells score	0.658	(0.511, 0.805)	92.31	39.29	58.54	84.62
	Revised Geneva score	0.624	(0.474, 0.773)	96.15	28.57	55.56	88.89
	AADD score	0.692	(0.550, 0.835)	88.46	50.00	62.16	82.35
	Years score	0.655	(0.508, 0.802)	84.62	46.43	59.46	76.47
	PEGeD score	0.673	(0.528, 0.818)	84.62	50.00	61.11	77.78
	4PEPS score	0.690	(0.546, 0.833)	80.77	57.14	63.64	76.19
Digestive system tumor	Wells score	0.549	(0.305, 0.792)	87.50	22.22	66.67	50.00
	Revised Geneva score	0.524	(0.282, 0.767)	93.75	11.11	65.22	50.00
	AADD score	0.549	(0.305, 0.792)	87.50	22.22	66.67	50.00
	Years score	0.517	(0.276, 0.759)	81.25	22.22	65.00	40.00
	PEGeD score	0.517	(0.276, 0.759)	81.25	22.22	65.00	40.00
	4PEPS score	0.549	(0.305, 0.792)	87.50	22.22	66.67	50.00
Reproductive system tumor	Wells score	0.688	(0.427, 0.948)	100.00	37.50	68.75	100.00
	Revised Geneva score	0.625	(0.356, 0.894)	100.00	25.00	64.71	100.00
	AADD score	0.659	(0.399, 0.919)	81.82	50.00	69.23	66.67
	Years score	0.767	(0.532, 1.000)	90.71	62.50	76.92	83.33
	PEGeD score	0.767	(0.532, 1.000)	90.91	62.50	76.92	83.33
	4PEPS score	0.631	(0.370, 0.891)	63.64	62.50	70.00	55.56

In addition, we compared the diagnostic value of five scores between the malignant and non-malignant tumor groups. The results presented in Table 6 indicated that, in the malignant tumor group, the Wells score, AADD score, Years score, PEGeD score, and 4PEPS demonstrated diagnostic superiority over the revised Geneva score (P < .05). In contrast, in the non-malignant tumor group, the AADD score, Years score, PEGeD score, and 4PEPS were diagnostically superior to both the Wells score and the revised Geneva score (P < .05).

Table 6.

The ROC Comparison Analysis of CDR Between the Scores Within the Cancer Group and Within the Non-cancer Group.

	Malignant Group		Non-Malignant Tumor Group
	z value	P value	z value	P value
Wells versus revised Geneva score	2.097	0.0360	1.067	0.2861
Wells versus AADD score	0.814	0.4154	3.459	0.0005
Wells versus Years score	0.717	0.4734	3.143	0.0017
Wells versus PEGeD score	0.991	0.3215	3.588	0.0003
Wells versus 4PEPS score	0.652	0.5147	2.345	0.0190
Revised Geneva versus AADD score	2.254	0.0242	3.366	0.0008
Revised Geneva versus Years score	2.256	0.0241	3.399	0.0007
Revised Geneva versus PEGeD score	2.463	0.0138	3.773	0.0002
Revised Geneva versus 4PEPS score	2.114	0.0345	2.731	0.0063
AADD versus Years score	0.037	0.9701	0.608	0.5431
AADD versus PEGeD score	0.352	0.7251	1.050	0.2935
AADD versus 4PEPS score	0.105	0.9160	0.126	0.8996
Years versus PEGeD score	1.000	0.3173	1.350	0.1770
Years versus 4PEPS score	0.070	0.9444	0.583	0.5600
PEGeD versus 4PEPS score	0.324	0.7463	1.140	0.2543

Discussion

The primary findings of this study indicated that when assessing the diagnostic value of CDR for PTE, various scoring criteria demonstrated significant differences between patients with malignant tumors and those without. In the malignant tumor group, the PEGeD score exhibited the highest overall predictive performance. The diagnostic performance of the Wells score, AADD score, Years score, PEGeD score, and 4PEPS was superior to that of the revised Geneva score; however, the AUC values for all these scores were below 0.7. In the non-malignant tumor group, the PEGeD score also exhibited the highest overall diagnostic performance. The AADD score, Years score, PEGeD score, and 4PEPS score all demonstrated superior diagnostic efficacy compared to the Wells score and the revised Geneva score, with AUC values exceeding 0.7 for all metrics except the revised Geneva score. Overall, the CDR has limited diagnostic value for patients with malignant tumors (AUC < 0.7) and was more appropriate for patients with non-malignant tumors.

The Wells score and the revised Geneva score are currently utilized to exclude PTE in the general population, as outlined in the international guidelines for the diagnosis of PTE.²⁰ However, in patients with malignant tumors, the Wells score is less effective for diagnostic purposes. Douma et al²¹ analyzed prospective data from a study that enrolled 3306 patients with suspected PTE, including 475 with malignancy. They found that the Wells score diagnosed PTE in patients with malignancy with an AUC of only 0.665 (with an AUC < 0.7 indicating poor model prediction), compared to 0.743 in patients with non-malignant tumors. In our study, we found that the AUC of the CDRs for diagnosing malignancy in conjunction with PTE was less than 0.7 for the Wells score, revised Geneva score, AADD score, Years score, PEGeD score, and 4PEPS. This indicated that the diagnostic efficacy of the aforementioned scoring criteria was suboptimal. A meta-analysis²² revealed a failure rate of 2.1% for the Wells criteria combined with D-dimer in predicting pulmonary embolism in patients with malignancy. Additionally, other revisions of the Geneva score, the AADD score, the Years score, and the PEGeD score all exhibited failure rates exceeding 2%.²² A failure rate greater than 2% indicated that these methods lacked reliable diagnostic value in patients at high risk for pulmonary embolism. In light of this, the 2022 ESMO guidelines²³ recommend that patients with suspected PTE associated with malignant tumors should undergo CTPA directly, rather than relying on CDR.

The scores included in this study had several shortcomings in the diagnosis of PTE in patients with malignant tumors. With the exception of the revised Geneva score, all scores included the subjective item, “alterative diagnosis less likely than PTE”. This item was a subjective measure that lacked clear criteria and relied heavily on the clinician's personal experience. Items in the Wells score can be diagnosed using an EKG, laboratory tests, chest X-rays, and other diagnostic methods.²⁴ Studies have shown that “alterative diagnosis less likely than PTE” serves as the primary criterion for differentiating between pulmonary embolisms (odds ratio: 3.7, 95% CI: 2.2-6.3).²¹ The “alterative diagnosis less likely than PTE” item in the 4PEPS was assigned a score of 5 points, making it more heavily weighted and subjective. In addition, we found that the percentage of hemoptysis and oxygen saturation levels below 95% in patients with malignant tumors was significantly higher than in those with non-malignant tumors. This was primarily due to the fact that a substantial proportion of malignant tumor patients were diagnosed with lung cancer, accounting for 40.6%. Furthermore, hemoptysis in lung cancer was often difficult to differentiate from that caused by pulmonary embolism. Additionally, intrapulmonary metastases and complications associated with malignant tumors may contribute to overall low oxygen saturation, which further decreased the predictive effect of the score.

D-dimer was a specific degradation product of cross-linked fibrin that is hydrolyzed by fibrinolytic enzymes. It was primarily used in clinical settings to exclude VTE. The use of D-dimer alone to diagnose pulmonary embolism in the general population had a sensitivity of approximately 95% and a specificity of around 50%, utilizing a cutoff value of 0.5 mg/L for fibrinogen equivalent units (FEU).²⁴ Of course, there were minor differences in diagnostic efficacy among various manufacturers of test reagents and testing methods. In addition, studies had shown that D-dimer levels were typically elevated in patients with malignant tumors. For instance, a study involving 290 cases of both benign and malignant breast tumors found that malignant breast tumors exhibited significantly higher mean D-dimer values (2837 ng/mL) compared to benign breast tumors (354 ng/mL).²⁵ The mean D-dimer level was significantly higher in the malignant group compared to the non-malignant group in this study, which aligned with existing literature. Schutgens RE et al²⁶ found that using D-dimer alone to diagnose PTE in patients with malignant tumors had a sensitivity of approximately 100%, but a specificity of only 21%. This low specificity rendered it unreliable for diagnosing patients at high risk of PTE. Regarding the causes of elevated D-dimer levels in patients with malignant tumors, it was important to consider not only the malignancies themselves but also other contributing factors. Patients with malignant tumors were particularly susceptible to influences such as infections, disseminated intravascular coagulation (DIC), cardiac failure, and pregnancy. Among these, elevated D-dimer levels due to infections were particularly common in this patient population.

It had been observed that the incidence of combined pneumonia in lung cancer patients can reach as high as 53%.²⁷ Additionally, the presence of concomitant risk factors such as tumor progression, chemotherapy, targeted therapy, and malnutrition increased the susceptibility of these patients to infections. This heightened vulnerability contributed to the poor specificity of D-dimer in diagnosing PTE in this population.²⁷ This was enhanced by adjusting the D-dimer cutoff value. Studies had shown that setting the D-dimer cutoff at 10 times the upper limit of normal (ULN) increased the specificity and positive predictive value for the diagnosis of VTE to over 90%.²⁸ The prediction of CDR can be significantly enhanced by appropriately increasing the cutoff value of D-dimer. One study utilized the Wells combined AADD score instead of relying solely on the D-dimer test, resulting in an increase in the percentage of patients with malignant tumors who were excluded from having PTE, rising from 6.3% to 12.6%.²⁹ Furthermore, some scholars have directly calculated the optimal cutoff value for D-dimer. It was found that, when using patients with a Wells score of ≤4 and suspected PTE as the study population, increasing the D-dimer cutoff value from 0.5 mg/L to 2.0 mg/L raised the specificity of the score from 2.7% to 21.7%.³⁰ The optimal D-dimer cutoff value varied significantly across different studies, highlighting the need for further research to determine the most appropriate D-dimer cutoff value for CDRs.

This is a retrospective study in which the study population was gathered from various departments. The level of detail in the patients’ historical questioning varied across departments, which may have led to the omission of certain data points. In addition, we conducted subgroup analyses of malignant tumors located in different sites. Although the Years score and PEGeD score exhibited some predictive value for reproductive system tumors (AUC = 0.767), their diagnostic utility was limited by the small sample size, which may have introduced bias into the experimental results. Due to the limited sample size, the scoring results for malignant tumors in other systems were not statistically significant, and comparisons between different scoring criteria were not feasible. In the future, we will increase the tumor sample size and conduct a systematic analysis of the predictive value of CDR for PTE patients with malignant tumors. With the rapid advancement of artificial intelligence technology, prediction models developed using this technology will possess significant advantages. In the future, it is anticipated that clinical decision rules for pulmonary thromboembolism, based on artificial intelligence, will be created to better serve patients with malignant tumors.

Conclusion

In summary, the Wells score, AADD score, Years score, PEGeD score, and 4PEPS were more effective than the revised Geneva score for the diagnosis of PTE in patients with malignant tumors. Overall, the CDRs demonstrated suboptimal diagnostic performance for PTE in patients with malignant tumors, while it appeared to be more effective for those with non-malignant tumors.

Footnotes

Acknowledgements

Not applicable.

Author Note

Jin Yang is also affiliated with Department of Pulmonary and Critical Care Medicine, Chen Zhou 3RD People's Hospital, Chen Zhou, Hunan Province, China.

Ethical Considerations

All procedures were approved by the ethics committee of Yantaishan Hospital (Ethic number: No.2025079). The investigations were carried out following the rules of the Declaration of Helsinki. The requirement for written informed consent was exempted because of the retrospective nature.

Author Contributions

Jin Yang: data curation, formal analysis, investigation, writing – original draft; Haixia Wang: methodology, visualization, writing – original draft; Jin Han: data curation, formal analysis, investigation, writing – original draft; Liang Zhang: project administration, supervision; Hongfu Ma: resources, supervision, validation; Xiaoxi Liu: data curation, formal analysis; Chunli Li: validation, investigation; Hao Chen: project administration, resources; Huimin Zhang: supervision, validation; Jianwen Fei: writing–review & editing. All authors reviewed the manuscript.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by grants from “Shandong Province Key Projects in Medical and Health Science and Technology—A study on constructing an AI-assisted VTE early warning model and a novel prevention system for high-risk hospitalized patients based on big data” (202403021092).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

Data will be made available on request.

ORCID iDs

Jin Yang

Haixia Wang

Jin Han

Chunli Li

Hao Chen

Huimin Zhang

Jianwen Fei

References

Lin

. Prevention, control, and standardized treatment of venous thromboembolism. Guangdong Med. 2022;43(3):265-270.

Kuan Wang

. Comparison of the diagnostic value of the YEARS score and age-adjusted D-dimer in patients with suspected pulmonary embolism. Mod Med Health Care. 2025;41(5):1111-1115.

Wells

Anderson

Rodger

, et al. Derivation of a simple clinical model to categorize patients probability of pulmonary embolism: increasing the models utility with the SimpliRED D-dimer. Thromb Haemost. 2000;83(3):416-420.

Le Gal

Righini

Roy

, et al. Prediction of pulmonary embolism in the emergency department: the revised Geneva score. Ann Intern Med. 2006;144(3):165-171.

Mulder

Horváth-Puhó

van Es

, et al. Venous thromboembolism in cancer patients: a population-based cohort study. Blood. 2021;137(14):1959-1969.

Posadas-Martínez

Vázquez

Giunta

Waisman

de Quirós

Gándara

. Performance of the wells score in patients with suspected pulmonary embolism during hospitalization: a delayed-type cross sectional study in a community hospital. Thromb Res. 2014;133(2):177-181.

Van

Douma

Mos

Huisman

Kamphuisen

. Performance of four clinical decision rules in patients with malignancy and suspected pulmonary embolism. J Thromb Haemost. 2012;10(2):312-314.

Gupta

Raja

Khorasani

. Assessing 2 D-dimer age-adjustment strategies to optimize computed tomographic use in ED evaluation of pulmonary embolism. Am J Emerg Med. 2014;32(12):1499-1502.

van der Hulle

Cheung

Kooij

, et al. Simplified diagnostic management of suspected pulmonary embolism (the YEARS study): a prospective, multicentre, cohort study. Lancet. 2017;390(10091):289-297.

10.

Kearon

de Wit

Parpia

, et al. Diagnosis of pulmonary embolism with d-dimer adjusted to clinical probability. N Engl J Med. 2019;381(22):2125-2134.

11.

Roy

Friou

Germeau

, et al. Derivation and validation of a 4-level clinical pretest probability score for suspected pulmonary embolism to safely decrease imaging testing. JAMA Cardiol. 2021;6(6):669-677.

12.

Nagel

Steffen

Schwartz

Hamm

Elgeti

. Age-dependent diagnostic accuracy of clinical scoring systems and D-dimer levels in the diagnosis of pulmonary embolism with computed tomography pulmonary angiography (CTPA). Eur Radiol. 2019;29(9):4563-4571.

13.

Eksioglu

Azapoglu Kaymak

Elhan

Cimilli Ozturk

. A comparative analysis of the impact of severe acute respiratory syndrome coronavirus 2 infection on the performance of clinical decision-making algorithms for pulmonary embolism. J Clin Med. 2024;13(23):7008.

14.

Vandenbroucke

von Elm

Altman

, et al. Strengthening the reporting of observational studies in epidemiology (STROBE): explanation and elaboration. Epidemiology. 2007;18(6):805-835.

15.

Konstantinides

Meyer

Becattini

, et al. 2019 ESC guidelines for the diagnosis and management of acute pulmonary embolism developed in collaboration with the European respiratory society (ERS). Eur Heart J. 2020;41(4):543-603.

16.

Kraaijpoel

van Es

Büller

, et al. The performance of the original and simplified wells scores in combination with age-adjusted D-dimer testing in the diagnostic management of pulmonary embolism. Blood. 2016;128(22):2569-2569.

17.

Speksnijder

Hessels

Muusses

Büller

Boersma

Simsek

. Validation of the YEARS algorithm and Wells’ score with the age-adjusted cut-off to exclude pulmonary embolism in COVID-19 patients. Thromb Res. 2022;219:109-111.

18.

Vielhauer

Benesch

Pernpruner

, et al. How to exclude pulmonary embolism in patients hospitalized with COVID-19: a comparison of predictive scores. Thromb J. 2023;21(1):51.

19.

Stals

MAM

Beenen

LFM

Coppens

, et al. Performance of the 4-level pulmonary embolism clinical probability score (4PEPS) in the diagnostic management of pulmonary embolism: an external validation study. Thromb Res. 2023;231:65-75.

20.

Linnemann

Blank

Doenst

, et al. Diagnostics and therapy of venous thrombosis and pulmonary embolism. The revised AWMF S2k guideline. Vasa. 2023;52(S111):1-146.

21.

Douma

van Sluis

Kamphuisen

, et al. Clinical decision rule and D-dimer have lower clinical utility to exclude pulmonary embolism in cancer patients. Explanations and potential ameliorations. Thromb Haemost. 2010;104(4):831-836.

22.

Stals

MAM

Takada

Kraaijpoel

, et al. Safety and efficiency of diagnostic strategies for ruling out pulmonary embolism in clinically relevant patient subgroups : a systematic review and individual-patient data meta-analysis. Ann Intern Med. 2022;175(2):244-255.

23.

Falanga

Nisio

, et al. Venous thromboembolism in cancer patients: ESMO clinical practice guideline. Ann Oncol. 2023;34(5):452-467.

24.

Gotta

Gruenewald

Eichler

, et al. Unveiling the diagnostic enigma of D-dimer testing in cancer patients: Current evidence and areas of application. Eur J Clin Invest. 2023;53(10):e14060.

25.

Stukan

Badocha

Ratajczak

. Development and validation of a model that includes two ultrasound parameters and the plasma D-dimer level for predicting malignancy in adnexal masses: an observational study. BMC Cancer. 2019;19(1):564.

26.

Schutgens

Esseboom

Haas

Nieuwenhuis

Biesma

. Usefulness of a semiquantitative D-dimer test for the exclusion of deep venous thrombosis in outpatients. Am J Med. 2002;112(8):617-621.

27.

Chinese Anti-Cancer Association Committee on Respiratory Oncology Chinese Anti-Cancer Association. Chinese Expert consensus on diagnosis, treatment and prevention of pulmonary infection in patients with solid tumors (2024 edition). Zhonghua Zhong Liu Za Zhi. 2024;46(12):1156-1175.

28.

Koch

Martin

Gruber-Rouh

, et al. Cancer patients with venous thromboembolism: diagnostic and prognostic value of elevated D-dimers. Eur J Clin Invest. 2023;53(4):e13914.

29.

Wilts

Gal

Den Exter

, et al. Performance of the age-adjusted cut-off for D-dimer in patients with cancer and suspected pulmonary embolism. Thromb Res. 2017;152:49-51.

30.

Kwon

Kim

Her

Chae

Lee

. Elevation of the D-dimer cut-off level might be applicable to rule out pulmonary embolism for active cancer patients in the emergency department. Intern Emerg Med. 2022;17(2):495-502.