Hyperspectral Imaging: Early Warning of Low Tissue Perfusion

The correct diagnosis can lead to early recognition and management of patient and wound care challenges.   

The EURODIALE study¹ reported that infection predicted nonhealing in individuals with diabetic foot ulcers (DFU), but only in subjects with peripheral arterial disease (PAD), making rapid awareness of PAD a high priority guiding care decisions and expected outcomes for DFU. Other risk factors for nonhealing, identified by logistic regression in this prospective cohort study of 1,088 subjects with a DFU, included larger DFU size, male gender, older age, inability to stand or walk without assistance, and prior heart failure or renal disease as independent baseline predictors of non-healing with or without amputation during 1 year of followup.   

Do the same findings apply to a patient with diabetes who has venous insufficiency? What if diabetes is removed from the equation? Does infection predict nonhealing for all those with a chronic wound only if its blood supply is compromised? We are only beginning to understand how interactions among different disease states such as PAD, venous insufficiency, and diabetes change the equations predicting whether or not an individual will proceed to healing. The 2 articles reviewed below highlight an early warning technique for identifying PAD and other danger signals indicating when a DFU may not be on the path to healing.

Reference: Nouvong A, Hoogwerf B, Mohler E, Davis B, Tajaddini A, Medenilla E. Evaluation of diabetic foot ulcer healing with hyperspectral imaging of oxyhemoglobin and deoxyhemoglobin. Diabetes Care. 2009;32(11):2056-2061.   

Rationale: Having a way to assess healing potential, including microvascular function, may help clinicians triage wounds more accurately, leading to earlier intervention with more aggressive therapies, thereby helping avoid infections and amputations. Hyperspectral imaging is a new noninvasive diagnostic tool to quantify and map microcirculatory changes in oxyhemoglobin (OHb) and deoxyhemoglobin (DOHb) near wounds.   

Objective: Test the accuracy and validity of hyperspectral tissue oxygenation mapping (HTOM) in evaluating the healing potential of DFU in individuals with Type 1 or 2 diabetes.   

Methods: Three United States-based outpatient wound centers conducted a prospective cohort trial, monitoring HTOM in lower limbs and at the DFU edge of 66 patients during 11 visits over 24 weeks. The subjects ranged in age from 25 years to 81 years, and presented with Type 1 or 2 diabetes. Clinical evaluation included age, sex, ethnicity/race, weight, height, BMI, HbA1c, systolic and diastolic blood pressure, ankle brachial index (ABI), diabetes type and duration, neuropathy, and transcutaneous oxygen tension (tcPO2) measured at the ankle of both legs. Superficial tissue oxygen saturation was measured in a 1 cm diameter area bordering the DFU, avoiding callus, and calculated as the percent of all hemoglobin that was oxygenated. Appropriate c² or t test statistics compared values of all variables measured for DFU that completely epithelized during the 24-week study to those of DFU that did not heal.   

Results: The only parameters statistically distinguishing the 38 patients who healed from the 16 who did not heal during the 24-week study were at the ulcer edge: OHb, superficial tissue oxygen saturation, and a healing index derived from the linear discriminant decision line of the plotted OHb and DOHb values. Among 73 DFU, 43(80%) of the 54 that healed were correctly predicted to heal (80% sensitivity). Of the 19 that failed to heal, 14 were predicted not to heal (74% specificity). Among the 48 initially predicted to heal, 43 did so (90% positive predictive validity). Of the 25 initially predicted not to heal, 15 were accurately predicted (60% negative predictive validity), increasing to 76% by excluding false negative predictions at 4 sites with callus. Excluding 3 sites with underlying osteomyelitis reduced false positive predictions of ulcer healing.   

Author’s conclusions: Hyperspectral imaging of tissue oxygenation can map tissue microvascular perfusion near a DFU on admission with spatial resolution of 100 microns without coming into contact with the patient’s foot. Hyperspectral tissue oxygenation mapping provides information necessary to treat and monitor foot complications faster and more specifically than would be possible with currently available methods.

Reference: Yekta Z, Pourali R, Nezhadrahim R, Ravanyar L, Ghasemi-rad M. Clinical and behavioral factors associated with management outcome in hospitalized patients with diabetic foot ulcer. Diabetes Metab Syndrome Obes Targets Ther. 2011;4:371-375.   

Rationale: Increasing age, male gender, prior DFU history, duration of diabetes, and related comorbidities, or complications such as hypertension, nephropathy, and retinopathy, are reported risk factors for amputation that remain to be confirmed across cultures and communities.   

Objective: Conduct a cross-sectional study to describe the prevalence of amputation and describe the demographic, clinical, and behavioral factors associated with amputation in hospitalized DFU patients in 2 medical training hospitals in Iran.   

Methods: A questionnaire administered to a random sample of patients with DFU hospitalized in each of the 2 training hospitals from September 2009 to December 2010 was used to collect standardized patient and DFU data. Variables recorded included the risk factors listed under Rationale above; educational level; marital status; diabetes-related laboratory values; behavioral factors such as smoking, or a body mass index (BMI) ≥ 25 kg/m²; Wagner Ulcer Classification Grade; length of hospitalization; type of therapy; and comorbidities, including ischemic heart disease and dyslipidemia.   

For each variable, the percent of subjects experiencing a major or minor amputation were compared to those without amputation using the c² test for categorical variables and independent Student’s t test for continuous variables. All variables were tested for association with amputation using a logistic regression model to assess the odds ratios of amputation compared to non-amputation, with a 95% confidence interval indicating statistical significance. Sensitivity and specificity of HbA1c and blood glucose levels at admission as predictors of amputation were calculated using a cutoff point derived from each variable’s receiver operating characteristic curve.   

Results: Most of the 94 patients sampled were male; overweight (BMI ≥ 25 kg/m2) with Type 2 diabetes; a Wagner Ulcer Classification Grade ≥ 3 DFU; and at least one complication of diabetes. Most had less than a high school education. The mean age was 60 years; mean BMI was 26.95. Amputation occurred in 32 patients (34%). Higher risk of amputation was associated with increased age ( > 50 years), less education, and being single. Gender did not predict amputation. Clinical factors associated with amputation were longer duration of diabetes; at least one complication of diabetes; prior DFU history; being overweight; admission HbA1c > 9.75 g/dL; blood glucose > 305 g/dL; and Wagner Grade > 3 (ie, deep ulcer with abscess or osteomyelitis). Using the combined HbA1c and blood glucose cutoff criteria predicted amputation with 81.5% sensitivity and 75.2% specificity.   

Author’s conclusions: Increasing patients’ and health care professionals’ knowledge of how to improve glycemic control and early DFU diagnosis may prevent the development of high Wagner-grade wounds. This would help ease the burden of DFU in Iran.  

Though the studies reviewed above^2,3 were on DFU, the findings may apply to other chronic wounds. Venous and pressure ulcers may also be complicated by impaired local circulation,⁴ and are less likely to recur,⁵ and more likely to improve,⁴ if patients or professionals are educated about, and use, effective patient and ulcer management techniques. Like DFU, greater initial area, depth, and severity strongly predict nonhealing in 6 months.^4,6   As Yekta et al³ note, their sample was from more serious hospitalized ulcers on patients with a HbA1c value ≥ 9.75, predicting amputation. Both healed and non-healed groups of outpatients managed by Nouvong et al² had average HbA1c values ≤ 9.7, below levels predicting a nonhealing amputation outcome in the hospitalized Iranians. This supports the relevance of HbA1c > 9.75 as an important predictor of a nonhealing and/or amputation.   If results by Nouvong et al² are replicated in larger studies, HTOM may offer a futuristic approach to point, click, and know if a patient’s ulcer has impaired perfusion. This information could help professionals focus on, and resolve, PAD issues before they reach the more serious Wagner Grade ≥ 3 status described by Yekta et al.³ Combined, these studies suggest that the local microvascular perfusion status of a DFU may be used early, before ulcer infection develops, to guide care toward earlier healing outcomes and away from amputation.

Laura Bolton, PhD, FAPWCA
Adjunct Associate Professor Department of Surgery,
UMDNJ WOUNDS

Editorial Advisory Board Member and Department Editor