Predicting complications and morbidities in PAD patients through lower extremity compositions with dual-energy CT and material decomposition: a 2-year follow-up observational study

Abstract

Background

Peripheral artery disease (PAD) is associated with various morbidities. This study aims to investigate the correlation between different lower extremity compositions and development of morbidities in PAD patients.

Methods

Between January 2018 and December 2020, 108 subjects diagnosed of PAD were enrolled (mean age of 64.1 ± 13.5 years) and utilized dual-energy computed tomography (DECT) with material decomposition to measure the vessel volume, muscle volume, fat volume, and cortical-bone volume in lower extremity respectively. The association between each leg composition and developing complications or morbidities in PAD patients was analyzed over a two-year follow-up.

Results

Fontaine stage 3 and 4 had lower muscle mass compared to stages 1 and 2. More severe vascular stenosis was associated with lower muscle, fat, and cortical-bone volume. Patients with severe Fontaine stages (3 and 4) and lower-leg vascular stenosis had a higher risk of developing infection or inflammation (OR 45.5, 95% CI: 13.5–166.7, and OR 11.7, 95% CI: 2.8–50, P < 0.05) and amputation (OR 18.2, 95% CI: 2.2–142.8, and OR 10.7, 95% CI: 1.11–100, P < 0.05). Lower thigh cortical-bone volume was associated with an increased risk of falls resulting in fractures (OR: 1.39, 95% CI: 1.13–2.19, P < 0.01). Thigh cortical-bone volume below 64.5 cm³ was identified as the cut-off value to predict fall-related fractures, with a sensitivity of 100% and specificity of 92%.

Conclusions

This study demonstrates the potential of DECT with material decomposition to assess lower extremity composition and its relevance in predicting complications and morbidities in PAD patients. Severe vascular stenosis may contribute to muscle wasting and subsequent complications, while lower thigh cortical-bone mass may serve as a predictor of fall-related fractures.

Background

Peripheral artery disease (PAD) occurs when stenosis or occlusion in the arterial lumen restricts blood flow to the lower extremities. PAD affects 3–10% of the global population [1], with more than 200 million diagnosed cases worldwide [2].

Symptoms of PAD include claudication, rest pain, inflammation, infection, tissue loss, and critical limb ischemia that can decrease the patient’s quality of life and lead to poor outcomes [3]. Previous studies have established a link between PAD and sarcopenia [4, 5]. Limb ischemia triggers reactive oxygen species production, inflammation, mitochondrial dysfunction, and tissue hypoperfusion, all of which can cause muscle atrophy and disability [6, 7]. Tsai et al. recently found that muscle wasting in the lower extremities is not only strongly associated with the severity of PAD but also occurs earlier than wasting of the abdominal muscles [8]. This suggests that compositions specific to the legs may be directly influenced by PAD, which may be associated with advanced complications or morbidity. However, the relationship between different compositions of the lower extremities, such as vessels, muscles, fat, as well as bone, and the development of complications or morbidity in patients with PAD remains unclear.

Dual-energy computed tomography (DECT) is a medical imaging technique that utilizes two different energy levels to produce images of the body, enabling the identification of material composition by collecting tissue characterization data through two absorption measurements employing photon spectra with high and low energy levels [9, 10]. Material decomposition can be used to differentiate tissues and objects, such as soft tissue, bone, and iodine-containing contrast agents, based on both their atomic number and their electron density. DECT has various medical applications, including detecting and characterizing different types of tumors [11,12,13,14,15,16] and assessing vascular disease [17,18,19,20,21], lung disease [22,23,24,25,26], stones [27, 28], and bone marrow edema [29].

The objective of this study was to explore the correlation between lower extremity composition and the development of complications or clinical morbidities in patients with PAD patients by using DECT-based material decomposition over a two-year follow-up period.

Methods

Study population and data collection

From January 2018 to December 2020, a total of 222 subjects were diagnosed with PAD via DECT of the legs at Mackay Memorial Hospital in Taipei, Taiwan. After excluding cases with poor-quality CT scan images, incomplete patient data and loss of follow-up over two-year period, 108 subjects were included in this study (Fig. 1). Of these subjects, 72 were male and 36 were female, with a mean age of 64.1 ± 13.5 years. Clinical data and CT images were retrospectively analyzed after de-identification by investigators. The clinical data included the occurrence of complications and clinical morbidities over a two-year follow-up period.

Fig. 1

Follow chat of study population selection. PAD: peripheral arterial disease; DECT: Dual-energy computed tomography

Baseline characteristics, including age, gender, height, body weight, and body mass index (BMI), were collected. The medical history of each patient was reviewed in detail, including known cardiovascular disease (CVD; including stroke or coronary artery disease), diabetes, hypertension (HTN), dyslipidemia, chronic kidney disease (CKD), and smoking habits. The composition of each leg, in terms of muscle volume, fat volume, cortical-bone volume, and iodine-containing contrast agent within the vessels (vessel volume), was determined (see “Derivation of the Material Decomposition” for details). Fontaine staging was performed to categorize patients into less-severe (stages 1 and 2) and severe (stages 3 and 4) groups for assessment of the clinical severity of stenosis in the lower extremity vasculature. For vessel measurement, we divided run-off arteries into three vascular segments, namely iliac, femoro-popliteal, and below-the-knee arteries. The scoring system was as described in previous study [8] and used to grade the severity in the lower extremities: 0, no stenosis; 1, < 50% stenosis; 2, ≥ 50% stenosis; and 3, total occlusion. The sum of scores obtained for each artery in the thigh or lower-leg was calculated. We divided the patients into “less severe” and “severe” groups according to the total scores (thigh: less severe, ≤ 3; severe, > 3; lower leg: less severe, ≤ 6; severe > 6).

We evaluated the following complications and morbidities including infection or inflammation of the lower extremities, onset of cardiovascular events (e.g., heart failure), falls with fractures, hospitalization for any cause, requiring vascular intervention for PAD, recurrence of PAD symptoms, and amputation of a lower extremity. The two-year follow-up period began after the CT scan on which the diagnosis was based. During the 2-year follow-up period, the study populations maintained the ongoing medical treatment regimen.

The Institutional Research Ethics Board (IRB) of Mackay Memorial Hospital approved the study (IRB number: 22MMHIS059e). The study was performed in accordance with the principles of the Helsinki Declaration.

DECT image collection and analysis

Contrast-enhanced DECT was performed on all patients by using a 256-slice dual-source CT scanner (SOMATOM® Definition Flash; Siemens Healthcare, Erlangen, Germany). The non-ionic contrast agents Omnipaque™ 300 (Iohexol, Daiichi-Sankyo, Tokyo, Japan) or Ultravist® 300 (Iopromide, Schering AG, Berlin, Germany) were intravenously administered using an automatic power injector at a rate of 3.0–4.0 mL/sec. A total of 80–100 mL of contrast agent was injected through the antecubital vein, followed by a 40 mL saline flush at the same rate. 10 s after injection, a contrast-enhanced DECT scan was conducted. Both high-energy (140kVp) and low-energy (80kVp) images were collected, with the scanning range extending from the lower abdomen (from the bony level of the L3-L4 vertebrae) to the plantar aspects of the feet. Patients were required to fast for at least 6 h prior to the examination. The scanning parameters were as follows: spiral scanning, dual-tube voltage of 140 and 80kVp, automatic tube current modulation of 200–230 mA, a pitch of 0.6, image reconstruction thickness of 2 mm, inter-reconstruction spacing of 1.5 mm, and matrix of 512 × 512 pixels (Table 1).

Table 1 Parameters of the dual-energy computed tomographic (DECT)

We divided the scanning field of interest into the thigh (femoral head to inferior femoral condyles) and lower-leg (whole tibia and fibula). We used a three-material decomposition method [30] to separate substances into the three categories of soft tissue (fat and muscle were designated as the same material in this study), cortical bone, and iodine-containing contrast agent in the vessel (Fig. 2). The regions of interest (ROIs) used in this study were defined with a radius of 10 pixels for fat measurement, 10 pixels for muscle measurement, and 5 pixels for cortical bone measurement.

Fig. 2

a Maximum intensity projection (MIP) images of run-off arteries and segments of the lower limbs. b Examples of different materials identified in this study. VB: volume fraction of cortical bone; Vs: volume fraction of soft tissue; V_VES: volume fraction of iodine contrast agent in the vessels. c 3D examples of different tissues. White: bone; beige: muscle; red: artery

Derivation of material decomposition

CT uses a continuous beam of X-rays passing through the human body at different angles to obtain projection images, which are then reconstructed via post-processing to generate a map of linear attenuation coefficients. The CT number (CTN) can be used to represent the relationship between the X-ray attenuation and each material in the CT image, and can be expressed as:

$${CTN}_{M}=[{\mu }_{M}-{\mu }_{W}/{\mu }_{W}] \times k=[{\mu }_{W}/{\mu }_{W}-1] \times k$$

(1)

where μ is the linear attenuation coefficient of a material, k is a constant, W is water, and M is the mixture of different specific materials.

By using the above equation, the linear attenuation coefficient of a mixture can be expressed as:

$${\mu }_{M}={\mu }_{W} [{CTN}_{M}/k +1]$$

(2)

The attenuation coefficient of a mixture can be defined as the sum of the attenuation coefficient of each material multiplied by its proportion in the mixture:

$${\mu }_{M} = {\text{V}}_{B} \times {\mu }_{B} + {V}_{S} \times {\mu }_{S} + {V}_{VES} \times {\mu }_{VES}$$

(3)

We defined a mixture, M as being composed of all three material categories (cortical bone, B; soft tissue, S; contrast medium in the vessels, VES). We used a volume-fraction formula, according to which the sum of the volumes of the materials equals the volume of the mixture, to express the volume percentages (V) of the three materials, as follows:

$$\left\{\begin{array}{c}{\text{V}}_{B} +{\text{V}}_{S} + {\text{V}}_{VES}=1\\ {\text{V}}_{B} = 1- {V}_{S} - {\text{V}}_{VES}\end{array}\right.$$

(4)

The CTN of a mixture at a certain energy level (kVp) is defined by the following formula:

$${CTN}_{M}={CTN}_{B} \times {\text{V}}_{B} + {CTN}_{S} \times {V}_{S} + {CTN}_{VES} \times {V}_{VES}$$

(5)

Therefore, the CTNs at high (h) and low (l) energy levels can be obtained as follows:

$$\left\{\begin{array}{c}{[{CTN}_{M}]}_{h}= {[{CTN}_{B} \times {V}_{B}]}_{h}+{[{CTN}_{S} \times {V}_{S}]}_{h} + {[{CTN}_{VES} \times {V}_{VES}]}_{h}\\ {[{CTN}_{M}]}_{\text{l}}= {[{CTN}_{B} \times {V}_{B}]}_{\text{l}}+{[{CTN}_{S} \times {V}_{S}]}_{\text{l}} + {[{CTN}_{VES} \times {V}_{VES}]}_{l}\end{array}\right.$$

(6)

Hence, by using Eq. (4), Eqs. (5) and (6) can be rewritten in the following forms:

$$\left\{\begin{array}{c}{CTN}_{M}={CTN}_{B} \times (1- {\text{V}}_{S}- {\text{V}}_{VES})+ {CTN}_{S} \times {\text{V}}_{S} + {CTN}_{VES} \times {\text{V}}_{VES}\\ {{[{CTN}_{M}]}_{h}}= {[{CTN}_{B} \times \left(1- {\text{V}}_{S}- {\text{V}}_{VES}\right)]}_{\text{h}}+{[{CTN}_{S} \times {V}_{S}]}_{\text{h}} + {[{CTN}_{VES} \times {V}_{VES}]}_{h}\\{[{CTN}_{M}]}_{l}= {[{CTN}_{B} \times (1- {\text{V}}_{S}- {\text{V}}_{VES})]}_{l}+{[{CTN}_{S} \times {\text{V}}_{S}]}_{l} + {[{CTN}_{VES} \times {\text{V}}_{VES}]}_{l}\end{array}\right.$$

(7)

$${\text{V}}_{B} =1- {\text{V}}_{S}- {\text{V}}_{VES}$$

(8)

According to volume conservation law, the volume percentages of three materials can be calculated using data acquired under two different energy levels. By solving a system of equations, the volume percentages of the three materials can be expressed as follows:

$$\left\{\begin{array}{c}{\text{V}}_{\text{S}}=\frac{{M}_{h}-{B}_{h} + ({VES}_{h}-{B}_{h})\times ({B}_{l}-{M}_{l})/({VES}_{l}-{B}_{l})}{{S}_{h}-{B}_{h}+\left({VES}_{h}-{B}_{h}\right)\times ({B}_{l}-{S}_{l})/({VES}_{l}-{B}_{l})}\\ {\text{ V}}_{\text{VES}}=\frac{{M}_{h}-{B}_{h} + ({S}_{h}-{B}_{h})\times ({B}_{l}-{M}_{l})/({S}_{l}-{B}_{l})}{{VES}_{h}-{B}_{h}+\left({S}_{h}-{B}_{h}\right)\times ({B}_{l}-{VES}_{l})/({S}_{l}-{B}_{l})}\\ {\text{V}}_{B} =1- {\text{V}}_{S}- {\text{V}}_{VES}\end{array}\right.$$

(9)

The values substituted on the right-hand side of the Eqs. (9) are the CTNs of the mixture and materials. The CTNs of 100% muscle and fat can be obtained by selecting regions of interest (ROIs) of muscle and fat from the dual-energy images of the patient. Thus, the CTNs of the three materials at high and low energies can be used to calculate their volume percentages.

Following three-material decomposition, we calculated the true volume of cortical bone, soft tissue, and contrast agent in the vessels of each patient by using the following equation:

$$\text{Total volume}=\Sigma\ \text{Volume fraction }\times \text{ voxel size}$$

Regarding the soft-tissue volume derived from material decomposition, we defined muscle as regions with a CT density of −29 HU to 150 HU and fat as those with a CT density of −190 HU to −30 HU. Accordingly, muscle and fat volumes were derived from the total volume of soft tissue via imaging segmentation in MATLAB, version: 9.13.0 (R2022b; MathWorks, Natick, MA, USA). The complete material decomposition workflow is presented in Fig. 3.

Fig. 3

Analysis flow chart of material decomposition using DECT

Measurement of Bone Mineral Density (BMD)

We used the CT Bone Mineral Analysis application on an IntelliSpace Portal (ISP) system (IntelliSpace Portal 12; Philips, Amsterdam, Netherlands) to analyze BMD, as described by Mueller et al. [31]. Soft-tissue and bone window setting images were used to visualize the images without affecting attenuation or BMD measurements. We focused on the ROIs in the axial cross-sections of the vertebral bodies of L4 and L5 for CT attenuation assessment. Using the ISP densitometry application, we calculated the average BMD and T scores (Fig. 4). To ensure accurate attenuation measurements, we avoided placing the ROIs in areas with attenuation heterogeneity, such as spinal hardware, compression fractures, spinal hemangiomas, and the posterior venous plexus. Finally, we collected the outcome variables for each patient.

Fig. 4

Example of bone mineral density measurement via CT

Statistical analysis

We present continuous variables as means ± standard deviations and categorical data as numbers and percentages. Statistical analysis was conducted using SPSS version 22.0 (IBM, Armonk, NY, USA). To assess between-group differences, analyses of variance, chi-squared tests and Fisher’s exact test were used. The Pearson product–moment correlation coefficient was utilized to determine correlations. We employed multivariable logistic regression to establish the odds ratio (OR) and 95% confidence interval (CI) of the contribution of vascular status as well as the composition of each leg composition (muscle, fat, and bone) to complications and morbidity after adjusting for age, gender, BMI, and clinical parameters (HTN, diabetes, dyslipidemia, CVD, CKD, and smoking habits). To minimize selection bias and the effects of confounders, propensity-score matching (PSM) was performed [32], and calculated using the risk factors determined by the abovementioned multivariable logistic regression modeling. One-to-one, nearest-neighbor matching without replacement and a caliper width of 0.2 were used to balance variables between the complication and non-complication groups. After matching, Mann–Whitney U and chi-squared tests were conducted to analyze group differences. Diagnostic performance was assessed using receiver operating characteristic (ROC) curve analysis, with the cut-off value calculated using the area under the ROC curve that yielded optimal sensitivity and specificity. A P value < 0.05 was considered statistically significant.

Results

The clinical features of the patients with PAD in this study (mean age: 64.1 ± 13.5 years, BMI: 25.5 ± 6.1 kg/m²) revealed a high incidence of HTN (74%), diabetes (67%), CVD (41%), CKD (54%), dyslipidemia (41%), and smoking (35%).

In our study, the volume of iodine-containing contrast agent within the vessels (vessel volume) was not correlated with the severity of vascular stenosis or the Fontaine stage. Patients with Fontaine stage 3 and 4 PAD had a lower muscle volume than that of patients with stage 1 and 2 PAD (thigh: 1769.9 ± 588.4 vs. 2066.8 ± 758.6 cm³; lower-leg: 1066.9 ± 325.9 vs. 1231.3 ± 394.6 cm³; all P < 0.05). According to the severity of upper-leg (thigh) or lower-leg vascular stenosis, the severe group had a lower muscle volume (thigh: 1585.6 ± 546.9 vs. 2297.6 ± 620.20 cm³; lower-leg: 986.8 ± 246.1 vs. 1271.8 ± 380.6 cm³), fat volume (thigh: 1344.9 ± 515.5 vs. 1679.7 ± 620.2 cm³; lower-leg: 401.0 ± 126.9 vs. 653.7 ± 303.0 cm³), and cortical-bone volume (thigh: 84.4 ± 24.9 vs. 98.8 ± 22.3 cm³; lower-leg: 80.2 ± 26.2 vs. 95.6 ± 28.2 cm³) than less-severe group (all P < 0.05). Fontaine stages 3 and 4 were also more prevalent in the severe vascular stenosis group, both in the thigh (73%) and the lower-leg (70%). Over the two-year follow-up period, patients who experienced infection/inflammation were more likely to experience muscle wasting at diagnosis (thigh: 1759.6 ± 575.3 vs. 2067.9 ± 719.2 cm³; lower-leg: 1061.7 ± 364.2 vs. 1222.9 ± 353.1 cm³; all P < 0.05), as did those who were hospitalized for any reason (thigh: 1779.9 ± 614.2 vs. 2194.8 ± 708.1 cm³; lower-leg: 1096.3 ± 357.2 vs. 1256.9 ± 355.8 cm³; all P < 0.05). Additionally, patients who underwent further vascular intervention had a lower fat volume (thigh: 1354.8 ± 487.3 vs. 1731.3 ± 785.3 cm³; lower-leg: 484.2 ± 194.9 vs. 606.7 ± 304.1 cm³; all P < 0.05) than those did not.

Patients who experienced falls resulting in fractures or hospitalization for any cause during the follow-up period had a lower cortical-bone volume at diagnosis in the thigh (51.9 ± 13.4 vs. 96.3 ± 22.3 cm³ and 88.2 ± 27.2 vs. 101.6 ± 16.5 cm³, respectively, all P < 0.05) and lower-leg (51.2 ± 6.6 vs. 93.3 ± 27.1 cm³ and 86.2 ± 29.5 vs. 97.4 ± 24.9 cm³, respectively, all P < 0.05) than those who did not experience. The details of complications and morbidities among patients with different leg compositions are listed in Table 2. The average BMD and T score of the lumbar spine did not differ between patients who experienced fractures or hospitalization and not. Neither was the average BMD or T score of the lumbar spine correlated with the thigh (r = 0.236 and 0.237, respectively) or lower-leg (r = 0.210 and 0.194, respectively) cortical-bone volume.

Table 2 Volumes (cm³) of leg tissues in relation to complications or morbidities

Patients with diabetes had lower cortical-bone and fat volumes in both the thigh and lower-leg, as well as a lower thigh muscle volume than those without diabetes. Those with CVD had a lower thigh muscle volume and cortical-bone volume in both the thigh and lower-leg, whereas individuals with dyslipidemia had a lower cortical-bone volume in the lower-leg than those without. No significant differences in leg composition were observed among the other medical condition groups. Detailed information on leg composition according to various medical conditions is available in Table 3.

Table 3 Volumes (cm³) of leg tissues in relation to medical conditions

Using multivariable logistic regression analysis, the factors associated with developing of complications and morbidities were evaluated. These factors included the volume of the thigh muscle, lower-leg muscle, thigh fat, lower-leg fat, thigh cortical-bone, lower-leg cortical-bone, Fontaine stage (1,2 vs. 3,4), and severity of vascular stenosis in thigh and lower-leg. We did not include the vessel volume in this analysis because it was not correlated with the severity of vascular stenosis or the Fontaine stage. The advanced Fontaine stage and severe lower-leg vascular stenosis were associated with a higher risk of infection/inflammation (OR: 45.5 [95% CI: 13.5–166.7] and OR: 11.7 [95% CI: 2.8–50.0], respectively, all P < 0.05) and amputation (OR: 18.2 [95% CI: 2.2–142.8] and OR: 10.7 [95% CI: 1.1–100.0], respectively, all P < 0.05) during the two-year follow-up. Additionally, thigh cortical-bone volume (per 1-cm³ decrement) was associated with an increased risk of falls resulting in fractures (OR: 1.39, 95% CI: 1.13–2.19, P < 0.01). We also applied Fisher’s exact test, which revealed that the low thigh cortical-bone volume group had a significantly higher risk of fracture (P < 0.05), while advanced Fontaine stage (3 and 4) was associated with a higher incidence of amputation (P < 0.01).

PSM was utilized to minimize selection bias in our retrospective study. The muscle volumes were lower in infection/inflammation group than those without (thigh: 1766.4 ± 590.5 vs. 2154.9 ± 794.4 cm³; lower-leg 1061.1 ± 385.2 vs. 1244.6 ± 343.6 cm³; all P < 0.05) before PSM. However, after PSM, these muscle volumes no longer significantly differed between the two groups. Nonetheless, owing to the limited number of fracture events among our study participants, PSM did not yield conclusive results for the link between leg cortical-bone volume and fractures. By ROC curve analysis, a thigh cortical-bone volume below 64.5 cm³ was identified as the cut-off value to predict fall-related fractures, with a sensitivity of 100% and specificity of 92%.

Discussion

DECT is a promising diagnostic imaging tool with numerous potential clinical applications. It provides valuable insights into the material properties of tissues and can be used to distinguish between tissues that have similar attenuation characteristics upon conventional single-energy imaging. DECT utilizes a three-material decomposition algorithm to create images specific to soft tissue, fat, and iodine-containing material. It provides accurate information on attenuation and iodine concentration [9, 33].

In this study, we found no significant correlation between the volume of iodine-containing vessels and the severity of vascular stenosis. The observed discordance in finding could be influenced by factors such as the presence of collateral circulation and the individualized size of the blood vessel lumen. However, we observed a significant reduction in lower extremity muscle volume in patients with Fontaine stages 3, 4 compared to that in patients with stages 1, 2. Tsai et al. [8] previously reported that muscle wasting in the lower extremities was more prominent and occurred earlier than wasting of the central muscles in PAD patients, and that the degree of muscle wasting corresponded to the severity of vascular stenosis. These findings all suggest that the degree of regional muscle wasting is proportional to the severity of vascular stenosis. We also observed that patients with more severe vascular stenosis had lower muscle, fat, and cortical-bone volumes than those with less severe vascular stenosis. These results suggest that PAD leads directly to the wasting of all the main tissues in the legs.

During the analysis of complications or morbidities, we observed that individuals who developed infection/inflammation had a lower muscle mass than those did not. After PSM, muscle volume did not significantly differ between patients with infection/inflammation and those without. The logistic regression analysis also indicated that only individuals with Fontaine stages 3, 4 (as opposed to stages 1, 2) and more severe vascular stenosis of lower-legs had a high risk of developing infection or inflammation. This finding suggests that more severe vascular stenosis may not only result in more muscle wasting but also increasing the risk of infection/inflammation. Fontaine stages 3, 4 as well as more severe lower-leg vascular narrowing were associated with a higher risk of amputation, but no significant correlation was found between these factors and recurrent symptoms or further vascular intervention. In the EUCLID trial, the overall rate of major amputation was 1.6% among PAD patients, with rates of 8.4% in those with critical limb ischemia at baseline and 1.2% in those without [34]. These results are consistent with the findings of our amputation subgroup.

Historically, the Fontaine stages have been widely used to classify chronic lower extremity ischemia owing to their simplicity. However, they lack objective diagnostic criteria for ischemia and have been largely replaced by updated threatened-limb classification systems, such as the Wound, Ischemia, and foot Infection (WIfI) classification system [35, 36]. Nonetheless, our two-year follow-up study demonstrated that a more severe Fontaine stage was associated with a greater risk of lower limb infection/inflammation or amputation. Therefore, it may still be utilized for risk prediction in patients with PAD.

Interestingly, a decrease in thigh cortical-bone volume was associated with an increased risk of falls resulting in fractures in this study, whereas neither the T score nor BMD of the lumbar spine was linked to the risk of fractures. We also found no correlation between the T score or BMD of the lumbar spine and the lower limb cortical-bone volume. Previous research by Ohlsson et al. showed that decreasing cortical thickness and area were correlated with an increased risk of fracture [37]. Furthermore, cortical bone area has been observed to be predictive of fractures independently of areal BMD in older men [38]. Moreover, in our study, ROC curve analysis indicated that a thigh cortical-bone volume below 64.5 cm³ may serve as a predictor of fall-related fractures. These findings suggest that cortical-bone volume may be affected by local ischemia and serve as a predictor of fracture development independent of BMD. It is worth noting that the sample size for the fracture is too small, requiring further validation and the 100% sensitivity observed may be biased due to the limited sample size. However, in our study, we used a direct measurement method (material decomposition) to determine the entire cortical-bone volume, which is expected to be more accurate and relatively straightforward. We did not perform BMD measurements of the hip or lower limb, and further investigation is necessary to evaluate the relationship between BMD and lower limb cortical-bone volume.

Due to the observational nature of this study and the limited number of subjects, a low incidence of fractures and amputations may be expected. The imbalance in cohort size, particularly in fracture and amputation cases, could introduce bias and limit the generalizability of the findings to the broader population. To mitigate bias, we employed methods such as logistic regression analysis, propensity-score matching, and Fisher’s exact test.

This study was subject to several limitations that should be addressed in further research. First, it was a retrospective and cross-sectional study, which limits its ability to establish causality. Second, the study was focused on patients with PAD, and additional research in other populations is necessary to broaden the generalizability of the results. Third, the sample was relatively small, particularly in the subgroup with fractures, and the follow-up period was only two years. To validate the findings, future studies with larger samples and longer follow-up periods are warranted.

In summary, we conducted a preliminary study to investigate the correlation between leg tissue volumes and clinical outcomes over a 2-year period. Our findings indicate that a higher Fontaine stage and more severe vascular stenosis in the lower-leg are associated with an increased risk of lower extremity infection, inflammation, or amputation during follow-up. Additionally, individuals with lower thigh cortical-bone volume demonstrated a higher risk of fall-related fractures. Our study suggests that Fontaine stage remains a useful and simple tool for predicting the risk of lower extremity infection, inflammation, and amputation in clinical practice. For PAD patients with low thigh cortical-bone volume, fall prevention strategies may be crucial to reducing fracture risk and preventing further morbidity. However, to confirm these findings, further studies with larger sample sizes and longer follow-up periods are needed.