Abstract
Background
The similar visual appearance of high‐risk basal cell carcinoma (BCC) and cutaneous squamous cell carcinoma (cSCC) may cause confusion for diagnosis. High‐frequency ultrasound (HFUS) may provide additional intralesional information and thus help to distinguish them.
Method
In this retrospective study, we analyzed the clinical characteristics, HFUS grayscale, and color Doppler flow imaging (CDFI) features of pathologically confirmed high‐risk BCC and cSCC lesions (n = 65 vs n = 68). Subsequently, discrimination models based on the significant HFUS features were established.
Results
Between high‐risk BCC and cSCC lesions, the HFUS grayscale features of the lesion size (10.0 mm vs 17.4 mm), thickness (3.1 mm vs 5.9 mm), internal hyperechoic spots (80.0% vs 23.5%), and posterior acoustic shadowing (16.9% vs 66.2%) were statistically different (all p < 0.001). As for the CDFI features, high‐risk BCC lesions mainly appeared as pattern II (47.7%), while cSCC lesions mainly appeared as pattern III (66.2%). Based on the above five features, an optimal discrimination model was established with a sensitivity of 91.2%, a specificity of 87.7%, and an accuracy of 89.5%.
Conclusion
HFUS features, including size, thickness, internal hyperechoic spots, posterior acoustic shadowing, and Doppler vascularity pattern, are useful for differential diagnosis between high‐risk BCC and cSCC.
Keywords: basal cell carcinoma, dermatology, squamous cell carcinoma, ultrasound
1. INTRODUCTION
Basal cell carcinoma (BCC) and cutaneous squamous cell carcinoma (cSCC) are the most common skin cancers and have become a global health problem with increasing incidence in recent years. BCC and cSCC belong to nonmelanoma skin cancer (NMSC). 1 The most significant risk factors for NMSC include ultraviolet (UV) exposure, age, fair skin, and immunosuppression. In general, BCC and cSCC develop in areas of the skin exposed to UV radiation, such as the head/neck, face, and extremities. As a result, both of them can lead to local destruction and disfigurement, and can affect patients mentally.
Compared with BCC, cSCC has unique high‐risk features, including chronic inflammatory process, rapid growing pattern, poor differentiation, and lymphovascular invasion. Thus, cSCC requires more aggressive diagnosis and excision methods. 2 However, the visual presentations of cSCC are quite similar to those of some high‐risk BCC forms, including infiltrative, micronodular, sclerodermiform, morpheaform, and basosquamous subtypes. 3 They present with ulcerations, local destruction, and telangiectasia, resulting in difficult diagnosis and even misdiagnosis. 4 , 5 Misdiagnosis resulting from similar‐looking lesions may lead to delayed treatment, higher medical cost, and functional and esthetic problems for patients. Therefore, new diagnostic modalities with more diagnostic information are of critical importance for differentiating high‐risk BCC from cSCC.
Biopsy is widely considered the gold standard to differentiate high‐risk BCC from cSCC. However, it may miss the aggressive part of the tumor because it only evaluates a limited section. 6 In addition, biopsy is invasive, and fails to provide spatial information regarding the size, thickness, and infiltration depth of the lesion.
Many noninvasive and immediate techniques, such as dermatoscopy, optical coherence tomography (OCT), reflectance confocal microscopy (RCM), and high‐frequency ultrasound (HFUS), are increasingly being used in diagnosing skin cancers. Among them, dermatoscopy is the most frequently used technique for evaluating skin cancers. However, dermatoscopy only provides superficial information of the lesion and cannot explore the internal components. Thus, it may miss the infiltrative features that may be present only at the deeper, advancing margins of the tumor. 7 OCT and RCM provide high‐resolution but incomplete intralesional information due to their relative low penetration (depth ≤0.5 mm). 8 Therefore, the above‐mentioned techniques are difficult for differentiating high‐risk BCC from cSCC. As a noninvasive and real‐time imaging technique, HFUS can provide a higher penetration than OCT and RCM, achieving a reasonable balance between resolution and penetration. 8 In addition, color Doppler flow imaging (CDFI) is valuable for assessing intralesional vascularity. Thus, HFUS may be a reliable tool for differentiating the two similar in visual appearance but completely different skin cancers.
To the best of our knowledge, the value of HFUS in the differentiation between high‐risk BCC and cSCC has not been well evaluated. Therefore, we conducted this study to explore in this regard, and further establish a practical discrimination model.
2. MATERIALS AND METHODS
This retrospective study was approved by the Ethics Committee of the university hospital, and informed consent was waived. The study was performed in accordance with the Declaration of Helsinki for medical research.
2.1. Patients
First, we reviewed the HFUS database of the Shanghai Skin Disease Hospital from December 2016 to July 2020. There were 233 BCC patients and 132 cSCC patients confirmed by pathology. Then, we selected 69 high‐risk BCC patients and 73 cSCC patients, according to the following inclusion criteria: (a) with complete and high‐quality images of each lesion, (b) with definitive histopathological diagnosis of high‐risk BCC or cSCC, and (c) without treatment prior to the HFUS examination. Subsequently, four high‐risk BCC patients and five cSCC patients were excluded because of the following exclusion criteria: (a) not primary high‐risk BCC (n = 1) and cSCC (n = 2) or (b) high‐risk BCC (n = 3) and cSCC (n = 3) with other malignant skin tumors or multiple lesions. Ultimately, 65 high‐risk BCC patients and 68 cSCC patients were enrolled in the study. A flowchart of the selected patients is shown in Figure 1.
FIGURE 1.
Flowchart of the selection of high‐risk basal cell carcinoma and cutaneous squamous cell carcinoma patients
2.2. HFUS examinations
All HFUS examinations were performed by the same professional radiologist (L.‐H. G.) with more than 5 years of experience in performing dermatologic HFUS. The HFUS device is a My Lab™ Class C scanner (Esaote SpA; Genoa, Italy) equipped with a linear transducer (frequency range, 10–22 MHz). We used the “Resolution High” mode to evaluate lesions in this study. In this mode, the transducer frequency was fine‐tuned between 18–22 MHz automatically.
Before each examination, the radiologist assisted the patients in maintaining the body position so that the lesions were fully exposed. Subsequently, sufficient amounts of gel were applied and the transducer was placed vertically and gently on the surface of the lesion. During the examination, each lesion was completely scanned. The gain, depth, and focus were adjusted to display each lesion clearly. In addition, CDFI parameters were adjusted to suppress noise artifacts, and to display color Doppler flow signals clearly. After the examinations, all HFUS images were stored digitally and analyzed in the following step.
2.3. Data and image analysis
Clinical characteristics, including patient gender, age, and location of the lesion, were collected for further analysis. The UV‐exposed sites were categorized as the face, eyelids, eyebrows, periorbital region, nose, lips, chin, mandible, preauricular region, postauricular region, temple, ear, forehead, scalp, and neck. Non‐UV exposed sites were categorized as genitalia, trunk, and extremities (excluding hands, feet, nail units, and ankles).
HFUS features, including CDFI images, were reviewed by two radiologists (X.‐L. L. and Q. W.) with 5 years and 3 years of experience, respectively, in diagnosing skin diseases using HFUS. They were blinded to the pathological results and evaluated the images with consensus.
2.4. Image interpretation
We defined and evaluated the following grayscale and CDFI features based on the previous literature 9 , 10 : (a) lesion size (the maximum diameter), measured at the longest diameter in the maximum plane; (b) thickness, measured at the thickest part of the lesion; (c) shape, categorized as regular or irregular; (d) margin of lesion base, categorized as flat, ridgy, irregular, or infiltrated; (e) margin of the lesion surface, categorized as concave, flat, convex, or crumpled; (f) homogeneous echogenicity, categorized as yes or no; (g) intralesional hyperechoic spot density, defined as at least two spots in any plane with diameters between 0.1 and 1.0 mm, categorized as present or absent; (h) posterior acoustic shadowing, categorized as present or absent; (i) layer involvement, categorized as confined to epidermis, involving dermis, or involving subcutaneous tissue; (j) pseudopodia, defined as lesion margin with spiculated protuberances, categorized as present or absent; (k) suspicious for lymph node involvement, categorized as yes or no; and (l) Doppler vascularity pattern, defined by vascular morphology and color signal coverage rate, categorized as pattern I, II, and III (Table 1 and Figure 2). The coverage rate was expressed as the ratio of the area of intralesional blood flow signals to the area of the lesion.
TABLE 1.
Doppler vascularity pattern
| Doppler vascularity pattern | Morphology | Coverage rate a |
|---|---|---|
| I | No blood flow signals or only scattered spotty blood flow signals | No coverage |
| II | Continuous linear blood flow signals with clear distribution | I < II < 50% |
| III | Dense linear blood flow signals with unclear distribution | ≥50% |
Expressed as the ratio of the area of intralesional blood flow signals to the area of the lesion.
FIGURE 2.
Images obtained from an 85‐year‐old female patient with high‐risk basal cell carcinoma (BCC) (A), a 70‐year‐old female patient with high‐risk BCC (B), and a 79‐year‐old male patient with cutaneous squamous cell carcinoma (C); (A) pattern I, showing scattered spotty blood flow signals at the bottom of the lesion; (B) pattern II, showing slender arborizing blood flow signals, no more than half of the lesion; (C) pattern III, showing twisted and dense linear blood flow signals, more than half of the lesion
2.5. Statistical analysis
Statistical analysis was performed using the Statistical Package for the Social Sciences version 26.0 (SPSS Inc. Chicago, IL, USA). Normality was evaluated using the Shapiro–Welk test. If continuous variables were skewed distributions, they were described as medians and interquartile ranges (IQR 25th to 75th percentile) and were analyzed using the Wilcoxon rank sum test. The remaining clinical and HFUS characteristics belonging to categorical variables were compared using the chi‐squared test or Fisher's exact probability test. After univariate analysis, binary logistic regression analysis was performed for features with p < 0.05. The forward likelihood ratio (LR) method was used to select independent variables in discrimination models. Finally, discrimination models were built based on statistically significant predictors. The cut‐off value for the discrimination model was 0.5. Sensitivity, specificity, and accuracy were calculated.
3. RESULTS
3.1. Patient characteristics
A summary of the clinical data is presented in Table 2. Between high‐risk BCC and cSCC cohorts, there was no statistical difference in gender (p = 0.135), while there was statistically significant difference in age (p < 0.001). The age of cSCC patients (median age, 79.5 years; IQR, 68.5–86.0 years) was greater than that of high‐risk BCC patients (median age, 68.0 years; IQR, 61.0–75.5 years) (p < 0.001). Moreover, the number of cSCC lesions at the non‐UV exposed sites was higher than that of high‐risk BCC (20.6% vs 6.2%, p = 0.015).
TABLE 2.
Basic characteristics and clinical data of the patients
| High‐risk basal cell carcinoma | Cutaneous squamous cell carcinoma | ||
|---|---|---|---|
| Characteristics | (n = 65) | (n = 68) | p value |
| Gender | 0.135 | ||
| Male | 26 (40.0%) | 36 (52.9%) | |
| Female | 39 (60.0%) | 32 (47.1%) | |
| Age a | 68.0 (61.0–75.5) | 79.5 (68.5–86.0) | <0.001 b |
| Location | 0.015 b | ||
| UV‐exposed site | 61 (93.8%) | 54 (79.4%) | |
| Non‐UV exposed site | 4 (6.2%) | 14 (20.6%) |
Note: Data are number of patients or lesions unless indicated otherwise.
Expressed as median, with interquartile ranges in parentheses.
Statistically significant difference.
3.2. Grayscale and CDFI features
The grayscale and CDFI features of the high‐risk BCC and cSCC lesions are presented in Table 3. No significant difference was found between the two groups in terms of lesion shape, margin of lesion base, layer involvement, pseudopodia, and suspicion of lymph node involvement. The thickness of cSCC (median thickness, 5.9 mm; IQR, 3.5–8.6 mm) was larger than that of high‐risk BCC (median thickness, 3.1 mm; IQR, 2.5–4.3 mm), and the size of cSCC (median size, 17.4 mm; IQR, 12.6–24.0 mm) was larger than that of high‐risk BCC (median size, 10.0 mm; IQR, 8.2–14.4 mm) (both p < 0.001).
TABLE 3.
The HFUS features for high‐risk basal cell carcinoma (BCC) and cutaneous squamous cell carcinoma (cSCC)
| High‐frequency ultrasound features | High‐risk BCC | cSCC | p value |
|---|---|---|---|
| Lesion size a | 10.0 (8.2–14.4) | 17.4 (12.6–24.0) | <0.001 b |
| Lesion thickness a | 3.1 (2.5–4.3) | 5.9 (3.5–8.6) | <0.001 b |
| Lesion shape | 0.455 | ||
| Regular, n (%) | 29 (44.6%) | 26 (38.2%) | |
| Irregular, n (%) | 36 (55.4%) | 42 (61.8%) | |
| Margin of lesion surface | 0.015 b | ||
| Concave, n (%) | 5 (7.7%) | 10 (14.7%) | |
| Flat, n (%) | 8 (12.3%) | 1 (1.5%) | |
| Convex, n (%) | 40 (61.5%) | 36 (52.9%) | |
| Crumpled, n (%) | 12 (18.5%) | 21 (30.9%) | |
| Margin of lesion base | 0.418 | ||
| Flat, n (%) | 15 (23.1%) | 9 (13.2%) | |
| Ridgy, n (%) | 11 (16.9%) | 17 (25.0%) | |
| Irregular, n (%) | 26 (40.0%) | 28 (41.2%) | |
| Infiltrated, n (%) | 13 (20.0%) | 14 (20.6%) | |
| Homogeneous echogenicity | 0.001 b | ||
| Yes, n (%) | 26 (40.0%) | 47 (69.1%) | |
| No, n (%) | 39 (60.0%) | 21 (30.9%) | |
| Posterior acoustic shadowing | <0.001 b | ||
| Present, n (%) | 11 (16.9%) | 45 (66.2%) | |
| Absent, n (%) | 54 (83.1%) | 23 (33.8%) | |
| Hyperechoic spots | <0.001 b | ||
| Present, n (%) | 52 (80.0%) | 16 (23.5%) | |
| Absent, n (%) | 13 (20.0%) | 52 (76.5%) | |
| Layer involvement | 0.283 | ||
| Confining to epidermis, n (%) | 3 (4.6%) | 8 (11.8%) | |
| Involving dermis, n (%) | 46 (70.8%) | 47 (69.1%) | |
| Involving subcutaneous tissue, n (%) | 16 (24.6%) | 13 (19.1%) | |
| Pseudopodia | 0.622 | ||
| Present, n (%) | 13 (20.0%) | 16 (23.5%) | |
| Absent, n (%) | 52 (80.0%) | 52 (76.5%) | |
| Lymphatic metastasis | 1.000 | ||
| Yes, n (%) | 0 (0.0%) | 1 (1.5%) | |
| No, n (%) | 65 (100.0%) | 67 (98.5%) | |
| Doppler vascularity pattern | <0.001 b | ||
| I, n (%) | 20 (30.8%) | 18 (26.5%) | |
| II, n (%) | 31 (47.7%) | 5 (7.4%) | |
| III, n (%) | 14 (21.5%) | 45 (66.2%) |
Note: Data are number of patients or lesions unless indicated otherwise.
Expressed as median, with interquartile ranges in parentheses.
Indicates statistically significant difference.
The composition ratio of the margin of lesion surface was inconsistent between high‐risk BCC and cSCC (p = 0.015). The margin of high‐risk BCC lesion surface was categorized as convex (61.5%), crumpled (18.5%), flat (12.3%), and concave (7.7%). The margin of SCC lesion surface was categorized as convex (52.9%), crumpled (30.9%), concave (14.1%), and flat (1.5%). The above content is described in Table 3.
More cSCC lesions showed homogeneous echogenicity in comparison with high‐risk BCC (69.1% vs 40.0%, p = 0.001). The hyperechoic spots in high‐risk BCC lesions were significantly more than that in cSCC lesions (80.0% vs 23.5%, p < 0.001). In addition, 45 (66.2%) cSCC lesions showed posterior acoustic shadowing, while only 11 (16.9%) high‐risk BCC lesions showed posterior acoustic shadowing (p < 0.001).
Regarding the Doppler vascularity pattern on CDFI, the composition ratio was inconsistent between high‐risk BCC and cSCC (p < 0.001). Most high‐risk BCC lesions showed pattern II vascularity (n = 31, 47.7%), but few showed pattern III vascularity (n = 14, 21.5%). On the contrary, most cSCC lesions showed pattern III vascularity (n = 45, 66.2%), but few showed pattern II vascularity (n = 5, 7.4%). The rate of high‐risk BCC and cSCC lesions showing pattern I vascularity was similar (30.8% vs 26.5%). (Figure 3)
FIGURE 3.
Images obtained from the nasal dorsum of a 66‐year‐old male patient with high‐risk basal cell carcinoma (BCC) (A, B, and C); images obtained from the nasion of an 88‐year‐old male patient with cutaneous squamous cell carcinoma (cSCC) (D, E, and F); (A & D) two clinically similar nodular lesions with ulcerations on the surface; (B) the grayscale image of high‐risk BCC shows a heterogeneous hypoechoic lesion (arrows) with intralesional hyperechoic spots (circle); (C) the color Doppler flow imaging (CDFI) of a high‐risk BCC lesion (arrows) shows continuous linear blood flow signals, no more than half of the lesion area (pattern II); (E) the surface of the cSCC lesion (arrows) shows thick linear hyperecho (triangle) accompanied by posterior acoustic shadowing (pentacle); (F) the CDFI of the cSCC lesion (arrows) shows the twisted and dense linear blood flow signals, more than half of the lesion (pattern III)
3.3. Developing the discrimination model
Based on Table 3, HFUS features, including lesion size, thickness, surface, homogeneous echogenicity, posterior acoustic shadowing, and hyperechoic spots as indexes were input into a binary regression analysis to determine the independent predictors. After multivariate analysis, the following four independent predictors were determined: lesion size (odds ratio [OR]: 1.098, p = 0.020), thickness (OR: 1.342, p = 0.021), hyperechoic spots (if present) (OR: 0.085, p < 0.001), and posterior acoustic shadowing (if present) (OR: 4.329, p = 0.006) (Table 4).
TABLE 4.
Multivariate analysis in establishing discrimination Model I
| Parameters | B | SE | Odds ratios | 95% CI | p value |
|---|---|---|---|---|---|
| Lesion size | 0.094 | 0.04 | 1.098 | 1.015−1.188 | 0.020 a |
| Lesion thickness | 0.294 | 0.127 | 1.342 | 1.046−1.723 | 0.021 a |
| Hyperechoic spots | <0.001 a | ||||
| Present | 2.459 | 0.543 | 0.085 | 0.029−0.248 | |
| Absent | 0 | 1 | |||
| Posterior acoustic shadowing | 0.006 a | ||||
| Present | 1.465 | 0.537 | 4.329 | 1.511−12.399 | |
| Absent | 0 | 1 |
Note: Hyperechoic spots (absent) and posterior acoustic shadowing (absent) used to be reference types.
Abbreviations: CI, confidence interval; SE, standard error.
Indicates statistically significant difference.
With the significant features, a logistic regression equation (Model I) was established as follows:
| (1) |
where each β value represents the B values, as shown in Table 4. The sensitivity, specificity, and accuracy of Model I were 83.8%, 86.2%, and 85.0%, respectively. Furthermore, the Doppler vascularity pattern combined with the above four grayscale features were used to establish Model II, which had a higher sensitivity (91.2%), specificity (87.7%), and accuracy (89.5%) compared with Model I.
4. DISCUSSION
Excision is the most effective treatment for BCC and cSCC, including standard excision and Mohs micrographic surgery. 2 , 11 cSCC has a high rate of recurrence (3%–16%), 12 , 13 neurological invasion (7.6%), 14 or lymphatic metastasis (4%). 13 If these occurred, further interventions, including dissection, adjuvant radiotherapy, and systemic treatment, were considered. 11 In contrast, metastasis of BCC is extremely rare, with an incidence of approximately 0.0028%–0.55%. 15 Therefore, according to the National Comprehensive Cancer Network guidelines, the minimum excision margin for cSCC (4–6 mm) is larger than that for BCC (4 mm). 7 , 16 All of these indicate that the prognosis for cSCC is worse than that for BCC. However, high‐risk BCC, as a part of BCC, is difficult to distinguish from cSCC in terms of appearance. 4 , 17 Based on our results, HFUS has potential to differentiate high‐risk BCC from cSCC noninvasively.
Our results showed that patients with cSCC were older than patients with high‐risk BCC, which was consistent with a previous study. Firnhaber suggested that up to 20% of BCC tumors occurred in patients under 50 years of age, while cSCC tumors in patients under 50 years of age were uncommon. 18 Also, we found that both high‐risk BCC and cSCC lesions were mainly located at UV‐exposed sites, which was consistent with a previous study. 4 This is because the main carcinogen, UV light, can induce genetic mutations. 15
Assessing the risk factors, such as lesion size and thickness, is crucial for improving treatment regimens. Based on our results, cSCC lesions were larger and thicker than high‐risk BCC lesions. Similarly, a previous study demonstrated that most cSCC lesions were thicker than 2.0 mm, measured on histological samples, 13 while the thickness of high‐risk BCC lesions was no more than 2.01 mm. 19 This may be related to the more aggressive growth pattern of cSCC spreading by local infiltration and expansion. 18 Lesion size has been considered as a risk factor for the recurrence of BCC or cSCC. The size of the high‐risk BCC lesion (10.0 mm) in the present study was similar to the actual size reported by Kricker et al. (11.0 mm). 20 On the other hand, the size of the cSCC lesion (17.4 mm) measured using HFUS in our results was slightly larger than the actual size (15.0 mm) reported in a previous study. 21 Therefore, we believed that the errors of the HFUS measurements for both high‐risk BCC and cSCC are limited and negligible. In brief, HFUS is a reliable method to measure the size and thickness of the high‐risk BCC and cSCC lesions noninvasively.
The feature of intralesional hyperechoic spots was a significant differentiator to make a distinction between high‐risk BCC and cSCC. Moreover, several studies have reported that a hyperechoic spot is a useful imaging sign for the diagnosis of BCC, and the number of hyperechoic spots is significantly higher in high‐risk BCC. 22 Although the origin of hyperechoic spots in BCC is unknown, it is clear that hyperechoic spots usually do not present posterior acoustic shadowing. 22 However, in cSCC, a thickened hyperechoic line on the surface of the lesion accompanied with posterior acoustic shadowing is a common finding on HFUS. The thickened hyperechoic line and posterior acoustic shadowing are both associated with the keratotic scales or crusts on the lesion surface. 10
In addition, we found a phenomenon that high‐risk BCC lesions showed more heterogeneous echogenicity than cSCC lesions. For BCC lesions, the reason may be that some BCC lesions are composed of mixed histological subtypes. 23 On the other hand, for cSCC lesions, the acoustic shadowing behind the hyperkeratinization may obscure the internal details of the lesion. The lesion surface was easily affected by external factors and could not accurately reflect the internal manifestation of the lesion. Therefore, homogeneous echogenicity and lesion surface could not serve as reliable evidences to distinguish high‐risk BCC from cSCC.
Our study indicated that both high‐risk BCC and cSCC lesions had abundant blood flow signals. Similarly, Vega et al. proposed that the BCC lesion with higher vascularity indicated more aggressive subtypes. 24 On the other hand, Wortsman found that hypervascularity could be visualized in cSCC lesions. 25 Therefore, it is difficult to distinguish the two diseases based on the feature of blood flow abundance.
However, according to the defined Doppler vascularity pattern, the morphology and coverage rate of high‐risk BCC and cSCC differed. Pathologically, the growth of high‐risk BCC extends along the path of least resistance to the underlying dermis. 26 On dermatoscopy, white structureless zones with fine linear vessels are predictors of high‐risk BCC. 15 , 27 This indicates that the nourishing vessels of high‐risk BCC are fine and dispersed, consistent with Doppler vascularity pattern II. On the contrary, cSCC have an aggressive growth pattern that expands and infiltrates into the dermis by extending along the tissue planes, conduits, perichondrium, or periosteum. 26
Furthermore, microvessel density and vascular endothelial growth factor levels are higher in cSCC than in high‐risk BCC. 28 The high degree of vascularization provides more oxygen and nutrients for tumor growth, which explains the more aggressive biological behavior of cSCC. In addition, cSCC usually appears as curved and abundant hairpin, linear‐irregular, or glomerular blood vessels on dermoscopy. 27 Correspondingly, HFUS shows dense linear blood flow signals with unclear distribution over half of the lesion. Therefore, it is feasible to distinguish between high‐risk BCC and cSCC based on the details of blood flow morphology and coverage rate. It should be mentioned that posterior acoustic shadowing below the keratinization sometimes prevents blood flow from being clearly visible, especially for cSCC. Therefore, we excluded lesions with severe hyperkeratosis in order to obtain high‐quality images.
Recently, elastography has been proven useful in the differentiation of malignant and benign skin lesions, preoperative assessment of the depth and border of skin cancers, and the predication of lymph node involvement. 29 , 30 , 31 In theory, elastography is also potentially useful in quantitatively differentiating high‐risk BCC and cSCC. However, there were some limitations in the clinical practice of elastography. First, the elastographic evaluation in skin lesions with ulcers or hyperkeratosis is not accurate. 32 Second, lesions located in curved parts of the body, such as nose or fingers, were difficult to fit in the transducer and to further be evaluated elastographically. 29 Overall, elastography has potential value in evaluating high‐risk BCC and cSCC. Regrettably, our previous device did not have the function of elastography. We will conduct the study of distinguishing high‐risk BCC from cSCC on a new device and report results in a timely manner.
There are several limitations in this study. First, this was a single‐center study based on Asian populations. It is uncertain whether our discrimination model is applicable to other races. Thus, further studies with a larger sample size are needed. Second, all identified HFUS features are subjective; therefore, there may be observational biases in clinical practice. Third, most of our data concerned the lesion located on skin exposed to UV radiation. However, the characteristics of the non‐UV exposed site lesions have not been fully evaluated according to the present study. We will further collect data on non‐UV exposed lesions and conduct ultrasound image analysis. Finally, we mainly focused on the HFUS features; however, the combination of clinical features and HFUS features is needed to develop the diagnostic model in further studies.
In conclusion, HFUS can reveal the internal features of high‐risk BCC and cSCC, such as lesion size, thickness, hyperechoic spots, posterior acoustic shadowing, and Doppler vascularity pattern. These features can be used to build a discrimination model to differentiate high‐risk BCC from cSCC. Therefore, HFUS is valuable for differentiating high‐risk BCC from cSCC.
Chen Z‐T, Yan J‐N, Zhu A‐Q, Wang L‐F, Wang Q, Li L, et al. High‐frequency ultrasound for differentiation between high‐risk basal cell carcinoma and cutaneous squamous cell carcinoma. Skin Res Technol. 2022;28:410–418. 10.1111/srt.13121
Zi‐Tong Chen and Jian‐Na Yan contributed equally to this work.
Contributor Information
Le‐Hang Guo, Email: gopp1314@hotmail.com.
Xiao‐Long Li, Email: 15275388623@163.com.
REFERENCES
- 1. Zalaudek I, Kreusch J, Giacomel J, Ferrara G, Catricala C, Argenziano G. How to diagnose nonpigmented skin tumors: a review of vascular structures seen with dermoscopy: part II. Nonmelanocytic skin tumors. J Am Acad Dermatol. 2010;63(3):377–86. [DOI] [PubMed] [Google Scholar]
- 2. Mendez BM, Thornton JF. Current basal and squamous cell skin cancer management. Plast Reconstr Surg. 2018;142(3):373e–87e. [DOI] [PubMed] [Google Scholar]
- 3. Hernandez‐Ibanez C, Blazquez‐Sanchez N, Aguilar‐Bernier M, Funez‐Liebana R, Rivas‐Ruiz F, de Troya‐Martin M. Usefulness of high‐frequency ultrasound in the classification of histologic subtypes of primary basal cell carcinoma. Actas Dermosifiliogr. 2017;108(1):42–51. [DOI] [PubMed] [Google Scholar]
- 4. Neagu N, Lallas K, Maskalane J, Salijuma E, Papageorgiou C, Gkentsidi T, et al. Minimizing the dermatoscopic morphologic overlap between basal and squamous cell carcinoma: a retrospective analysis of initially misclassified tumours. J Eur Acad Dermatol Venereol. 2020;34:1999–2003. [DOI] [PubMed] [Google Scholar]
- 5. Ryu TH, Kye H, Choi JE, Ahn HH, Kye YC, Seo SH. Features causing confusion between basal cell carcinoma and squamous cell carcinoma in clinical diagnosis. Ann Dermatol. 2018;30(1):64–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Wolberink EA, Pasch MC, Zeiler M, van Erp PE, Gerritsen MJ. High discordance between punch biopsy and excision in establishing basal cell carcinoma subtype: analysis of 500 cases. J Eur Acad Dermatol Venereol. 2013;27(8):985–9. [DOI] [PubMed] [Google Scholar]
- 7. Bichakjian CK, Olencki T, Aasi SZ, Alam M, Andersen JS, Berg D, et al. Basal cell skin cancer, version 1.2016, NCCN clinical practice guidelines in oncology. J Natl Compr Canc Netw. 2016;14(5):574–97. [DOI] [PubMed] [Google Scholar]
- 8. Wortsman X. Common applications of dermatologic sonography. J Ultrasound Med. 2012;31(1):97–111. [DOI] [PubMed] [Google Scholar]
- 9. Li M‐X, Wang Q, Li X‐L, Zhao C‐K, Zhu R‐Z, Chen J, et al. Imaging findings of Bowen's disease: a comparison between ultrasound biomicroscopy and conventional high‐frequency ultrasound. Skin Res Technol. 2020;26(5):654–63. [DOI] [PubMed] [Google Scholar]
- 10. Zhu A‐Q, Wang L‐F, Li X‐L, Wang Q, Li M‐X, Ma Y‐Y, et al. High‐frequency ultrasound in the diagnosis of the spectrum of cutaneous squamous cell carcinoma: noninvasively distinguishing actinic keratosis, Bowen's disease, and invasive squamous cell carcinoma. Skin Res Technol. 2021;27:831–40. [DOI] [PubMed] [Google Scholar]
- 11. Combalia A, Carrera C. Squamous cell carcinoma: an update on diagnosis and treatment. Dermatol Pract Concept. 2020;10(3):e2020066. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Eroglu A, Berberoglu U, Berreroglu S. Risk factors related to locoregional recurrence in squamous cell carcinoma of the skin. J Surg Oncol. 1996;61(2):124–30. [DOI] [PubMed] [Google Scholar]
- 13. Brantsch KD, Meisner C, Schönfisch B, Trilling B, Wehner‐Caroli J, Röcken M, et al. Analysis of risk factors determining prognosis of cutaneous squamous‐cell carcinoma: a prospective study. Lancet Oncol. 2008;9(8):713–20. [DOI] [PubMed] [Google Scholar]
- 14. Khan K, Mykula R, Kerstein R, Rabey N, Bragg T, Crick A, et al. A 5‐year follow‐up study of 633 cutaneous SCC excisions: rates of local recurrence and lymph node metastasis. J Plast Reconstr Aesthet Surg. 2018;71(8):1153–8. [DOI] [PubMed] [Google Scholar]
- 15. Peris K, Fargnoli MC, Garbe C, Kaufmann R, Bastholt L, Basset Seguin N, et al. Diagnosis and treatment of basal cell carcinoma: European consensus‐based interdisciplinary guidelines. Eur J Cancer. 2019;118:10–34. [DOI] [PubMed] [Google Scholar]
- 16. Christopher K, Sumaira Z, Alam M. National Comprehensive Cancer Network. Squamous Cell Skin Cancer (Version 1. 2019). https://www.nccn.org/professionals/physician_gls/pdf/squamous.pdf
- 17. Lombardi M, Pampena R, Borsari S, Bombonato C, Benati E, Pellacani G, et al. Dermoscopic features of basal cell carcinoma on the lower limbs: a chameleon! Dermatology. 2017;233(6):482–8. [DOI] [PubMed] [Google Scholar]
- 18. Firnhaber JM. Diagnosis and treatment of basal cell and squamous cell carcinoma. Am Fam Physician. 2012;86(2):161–8. [PubMed] [Google Scholar]
- 19. Welsch MJ, Troiani BM, Hale L, DelTondo J, Helm KF, Clarke LE. Basal cell carcinoma characteristics as predictors of depth of invasion. J Am Acad Dermatol. 2012;67(1):47–53. [DOI] [PubMed] [Google Scholar]
- 20. Kricker A, Armstrong B, Hansen V, Watson A, Singh‐Khaira G, Lecathelinais C, et al. Basal cell carcinoma and squamous cell carcinoma growth rates and determinants of size in community patients. J Am Acad Dermatol. 2014;70(3):456–64. [DOI] [PubMed] [Google Scholar]
- 21. Jambusaria‐Pahlajani A, Kanetsky PA, Karia PS, Hwang W‐T, Gelfand JM, Whalen FM, et al. Evaluation of AJCC tumor staging for cutaneous squamous cell carcinoma and a proposed alternative tumor staging system. JAMA Dermatol. 2013;149(4):402–10. [DOI] [PubMed] [Google Scholar]
- 22. Wortsman X, Vergara P, Castro A, Saavedra D, Bobadilla F, Sazunic I, et al. Ultrasound as predictor of histologic subtypes linked to recurrence in basal cell carcinoma of the skin. J Eur Acad Dermatol Venereol. 2015;29(4):702–7. [DOI] [PubMed] [Google Scholar]
- 23. Roozeboom MH, Mosterd K, Winnepenninckx VJ, Nelemans PJ, Kelleners‐Smeets NW. Agreement between histological subtype on punch biopsy and surgical excision in primary basal cell carcinoma. J Eur Acad Dermatol Venereol. 2013;27(7):894–8. [DOI] [PubMed] [Google Scholar]
- 24. Vega N, Wortsman X, Navarrete N, Sazunic I. Color Doppler ultrasound supports early diagnosis of mixed high and low risk of recurrence subtypes in the same basal cell carcinoma lesion. Dermatol Surg. 2018;44(5):741–3. [DOI] [PubMed] [Google Scholar]
- 25. Wortsman X. Ultrasound in dermatology: why, how, and when? Semin Ultrasound CT MR. 2013;34(3):177–95. [DOI] [PubMed] [Google Scholar]
- 26. Vargo N. Basal cell and squamous cell carcinoma. Semin Oncol Nurs. 2003;19(1):12–21. [DOI] [PubMed] [Google Scholar]
- 27. Lallas A, Argenziano G, Zendri E, Moscarella E, Longo C, Grenzi L, et al. Update on non‐melanoma skin cancer and the value of dermoscopy in its diagnosis and treatment monitoring. Expert Rev Anticancer Ther. 2013;13(5):541–58. [DOI] [PubMed] [Google Scholar]
- 28. Loggini B, Boldrini L, Gisfredi S, Ursino S, Camacci T, De Jeso K, et al. CD34 microvessel density and VEGF expression in basal and squamous cell carcinoma. Pathol Res Pract. 2003;199(11):705–12. [DOI] [PubMed] [Google Scholar]
- 29. Dasgeb B, Morris MA, Mehregan D, Siegel EL. Quantified ultrasound elastography in the assessment of cutaneous carcinoma. Br J Radiol. 2015;88(1054):20150344. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Unal M, Kaya Islamoglu ZG, Ozturk M, Uysal E, Islamoglu N. Can shear wave ultrasound elastography be a useful tool for determining the tumor extent of basal cell carcinomas pre‐operatively? A preliminary study. Skin Res Technol. 2020;26(1):148–9. [DOI] [PubMed] [Google Scholar]
- 31. Onol S, Ozkaya O. Diagnostic value of real‐time elastography in diagnosing lymph node metastasis of skin cancer. Cureus. 2020;12(10):e10997. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Tanaka T, Tada Y, Ohnishi T, Watanabe S. Usefulness of real‐time tissue elastography for detecting the border of basal cell carcinomas. J Dermatol. 2017;44(4):438–43. [DOI] [PubMed] [Google Scholar]