Dynamic analysis of EUS used for the differentiation of benign and malignant lymph nodes

EUS elastography for the differential diagnosis of pancreatic masses

AIM: To investigate the usefulness of endoscopic ultrasound (EUS) elastography for the differential diagnosis of malignant and benign pancreatic masses.

METHODS: Patients with pancreatic masses were enrolled in this study. All patients were examined by EUS and EUS elastography. Elastographic pattern was scored with whole numbers from 1 to 5 based on the color of elastographic images. Scores 1 and 2 were classified as benign and scores 3 to 5 as malignant.

RESULTS: Elastography was successfully performed in all patients and there were no procedure-related complications. Nineteen patients were finally diagnosed with pancreatic carcinoma, 1 with malignant pancreatic neuroendocrine tumor, 1 with benign endocrine tumor and 6 with focal pancreatitis. According to our diagnostic criteria of elastography, the overall accuracy of EUS elastography for the differential diagnosis of focal pancreatic masses was 88.89% (24/27), with a sensitivity, specificity, positive predictive value and negative predictive value of 100%, 57.14%, 86.96% and 100%, respectively.

CONCLUSION: EUS elastography has a relatively high accuracy for the differentiation of benign and malignant pancreatic lesions and may provide an alternative method when EUS-FNA reveals negative results or cannot be performed because of technical problems.

Efficacy of an artificial neural network-based approach to endoscopic ultrasound elastography in diagnosis of focal pancreatic masses

BACKGROUND & AIMS

By using strain assessment, real-time endoscopic ultrasound (EUS) elastography provides additional information about a lesion's characteristics in the pancreas. We assessed the accuracy of real-time EUS elastography in focal pancreatic lesions using computer-aided diagnosis by artificial neural network analysis.

METHODS

We performed a prospective, blinded, multicentric study at of 258 patients (774 recordings from EUS elastography) who were diagnosed with chronic pancreatitis (n = 47) or pancreatic adenocarcinoma (n = 211) from 13 tertiary academic medical centers in Europe (the European EUS Elastography Multicentric Study Group). We used postprocessing software analysis to compute individual frames of elastography movies recorded by retrieving hue histogram data from a dynamic sequence of EUS elastography into a numeric matrix. The data then were analyzed in an extended neural network analysis, to automatically differentiate benign from malignant patterns.

RESULTS

The neural computing approach had 91.14% training accuracy (95% confidence interval [CI], 89.87%-92.42%) and 84.27% testing accuracy (95% CI, 83.09%-85.44%). These results were obtained using the 10-fold cross-validation technique. The statistical analysis of the classification process showed a sensitivity of 87.59%, a specificity of 82.94%, a positive predictive value of 96.25%, and a negative predictive value of 57.22%. Moreover, the corresponding area under the receiver operating characteristic curve was 0.94 (95% CI, 0.91%-0.97%), which was significantly higher than the values obtained by simple mean hue histogram analysis, for which the area under the receiver operating characteristic was 0.85.

CONCLUSIONS

Use of the artificial intelligence methodology via artificial neural networks supports the medical decision process, providing fast and accurate diagnoses.

Predictors of malignancy in EUS-guided FNA for mediastinal lymphadenopathy in patients without history of lung cancer

BACKGROUND:

Mediastinal lymphadenopathy (ML) poses a great diagnostic challenge.

OBJECTIVE:

To investigate the predictors of malignancy in endoscopic ultrasound (EUS)-guided fine-needle aspiration (FNA) of ML in patients without known lung cancer.

DESIGN:

Retrospective study.

SETTING:

Tertiary referral center.

METHODS:

One hundred eight patients without known lung cancer who underwent EUS guided-FNA for ML between 2000 and 2007. All subjects underwent EUS-guided FNA. Data was collected on patients′ demographics, and lymph node (LN) characteristics. Diagnosis of LN malignancy was based on FNA findings and clinical follow-up.

RESULTS:

One hundred eight patients were analyzed; 58 (54%) were men and 87 (79%) were Caucasian. Mean age was 55 years. Prior malignancy was present in 48 (43%) patients. A total of 126 FNA samples from 126 distinct LNs were performed. Twenty-five (20%) LNs were positive for malignancy. Mean short and long-axis for LNs were 13 and 29 mms respectively. Round shape and sharp borders were found in 29 (15%) and 25 (22%) LNs, correspondingly. Independent predictors of a malignant FNA were: Prior cancer (OR 13.10; 95% CI 2.7-63.32; P = 0.001), short axis (OR 1.10; 95% CI 1.00-1.22; P = 0.041) and sharp LN borders (OR 5.47; 95% CI 1.01-29.51; P = 0.048). Age, race, gender, long axis, round shape were not associated with cancer in our cohort.

LIMITATIONS:

Retrospective design and lack of surgical gold standard.

CONCLUSIONS:

Increased risk of malignancy was associated with prior history of cancer, larger LN short axis and presence of LN sharp borders. These predictors may help guide endoscopists perform FNA in malignant LNs, increasing the overall efficiency of EUS-FNA for ML.

EUS elastography for the differentiation of benign and malignant lymph nodes: a meta-analysis

BACKGROUND

EUS elastography is a new technique for differentiating benign and malignant lymph nodes (LNs) by describing the mechanical property of the target tissue.

OBJECTIVE

To assess the accuracy of EUS elastography by pooling data of existing trials.

DESIGN

Seven studies involving 368 patients with 431 LNs were included. Meta-analysis was performed. Pooling was conducted in a fixed-effect model or a random-effect model.

PATIENTS

This study involved 368 patients.

INTERVENTION

EUS elastography.

MAIN OUTCOME MEASUREMENTS

Meta-analysis and meta-regression analysis.

RESULTS

The pooled sensitivity of EUS elastography for the differential diagnosis of benign and malignant LNs was 88% (95% confidence interval [CI] 0.83-0.92), and the specificity was 85% (95% CI, 0.79-0.89). The area under the curve under summary receiver operating characteristic (SROC) was 0.9456. The pooled positive likelihood ratio was 5.68 (95% CI, 2.86-11.28), and the negative likelihood ratio was 0.15 (95% CI, 0.10-0.21). The subgroup analysis by excluding the outliers provided a sensitivity of 85% (95% CI, 0.79-0.90) and a specificity of 91% (95% CI, 0.85-0.95) for the differential diagnosis of benign and malignant LNs. The area under the curve under SROC was 0.9421.

LIMITATIONS

A small number of studies met inclusion criteria.

CONCLUSION

EUS elastography is a promising, noninvasive method for differential diagnosis of malignant LNs and may prove to be a valuable supplemental method to EUS-guided FNA.

Quantitative measurement for thyroid cancer characterization based on elastography

OBJECTIVES

The purpose of this study was to evaluate color thyroid elastograms quantitatively and objectively and select more effective features to differentiate benign from malignant thyroid nodules.

METHODS

The study was approved by the Ethics Committee of Harbin Medical University. A total of 125 cases (56 malignant and 69 benign) were analyzed in this retrospective study. The original color thyroid elastograms were transferred from the red-green-blue color space to the hue-saturation-value color space. The elasticity information was represented by the hue component of color elastograms. The lesion regions were delineated by radiologists, and statistical and textural features were extracted. Then the most effective and reliable features among them were selected by using a minimum redundancy-maximum relevance algorithm. The selected features were input to a support vector machine to differentiate benign from malignant thyroid nodules.

RESULTS

The classification accuracy was 93.6% when the hard area ratio and textural feature (energy) of the lesion region were used. The area under the receiver operating characteristic curve for the hard area ratio was higher than that for the strain ratio (0.97 versus 0.87; P < .01), and the area under the curve for the hard area ratio was also higher than that for the color score (0.97 versus 0.80; P < .001). The results also showed that the features were robust for lesion region delineation.

CONCLUSIONS

The hard area ratio is an important and quantitative metric for elastograms. Quantitative analysis of elastograms using computer-aided diagnostic techniques can improve diagnostic accuracy.

Beyond conventional endoscopic ultrasound: elastography, contrast enhancement and hybrid techniques

PURPOSE OF REVIEW

Endoscopic ultrasound (EUS) recently became a technique with a major clinical impact in digestive diseases. EUS determines a change in the diagnosis and management of more than half of examined patients. This review summarizes recent advances in the complementary EUS examination modalities such as elastography and microbubble contrast enhancement.

RECENT FINDINGS

EUS elastography is a well documented method that allows characterization and differentiation of pancreatic cancer and chronic pancreatitis. Quantitative elastography methods, especially based on computer analyses, retrieve numeric values and possibly eliminate the human bias. The use of contrast-enhanced EUS also allows a better visualization and differentiation of focal pancreatic lesions. A hypoenhanced mass as compared with the surrounding pancreatic parenchyma is highly suggestive for pancreatic adenocarcinoma, whereas a hyperenhanced lesion indicates an inflammatory mass. Furthermore, hybrid EUS imaging techniques (in combination with computed tomography or magnetic resonance) might be useful for an increased diagnostic confidence.

SUMMARY

Despite its advantages in assessing the organs situated near the gastrointestinal tract, EUS is still an operator-dependent technique. The new EUS examination modalities incorporated in modern ultrasound systems allow a highly accurate diagnosis.

Ultrasound elastography as an adjuvant to conventional ultrasound in the preoperative assessment of axillary lymph nodes in suspected breast cancer: a pilot study

AIMS

To compare the performance of ultrasound elastography with conventional ultrasound in the assessment of axillary lymph nodes in suspected breast cancer and whether ultrasound elastography as an adjunct to conventional ultrasound can increase the sensitivity of conventional ultrasound used alone.

MATERIALS AND METHODS

Fifty symptomatic women with a sonographic suspicion for breast cancer underwent ultrasound elastography of the ipsilateral axilla concurrent with conventional ultrasound being performed as part of triple assessment. Elastograms were visually scored, strain measurements calculated and node area and perimeter measurements taken. Theoretical biopsy cut points were selected. The sensitivity, specificity, positive predictive value (PPV), and negative predictive values (NPV) were calculated and receiver operating characteristic (ROC) analysis was performed and compared for elastograms and conventional ultrasound images with surgical histology as the reference standard.

RESULTS

The mean age of the women was 57 years. Twenty-nine out of 50 of the nodes were histologically negative on surgical histology and 21 were positive. The sensitivity, specificity, PPV, and NPV for conventional ultrasound were 76, 78, 70, and 81%, respectively; 90, 86, 83, and 93%, respectively, for visual ultrasound elastography; and for strain scoring, 100, 48, 58 and 100%, respectively. There was no significant difference between any of the node measurements

CONCLUSIONS

Initial experience with ultrasound elastography of axillary lymph nodes, showed that it is more sensitive than conventional ultrasound in detecting abnormal nodes in the axilla in cases of suspected breast cancer. The specificity remained acceptable and ultrasound elastography used as an adjunct to conventional ultrasound has the potential to improve the performance of conventional ultrasound alone.

EUS elastography combined with the strain ratio of tissue elasticity for diagnosis of solid pancreatic masses

BACKGROUND

Recently, the usefulness of endoscopic ultrasound (EUS) elastography has been reported for the diagnosis of pancreatic lesions. In the present study, we retrospectively assessed EUS elastography as a diagnostic tool by evaluating tissue elasticity distribution and elasticity semiquantification, using the strain ratio (SR) of tissue elasticity, in patients with pancreatic masses.

METHODS

One hundred and nine patients who underwent EUS elastography between September 2006 and May 2009 were retrospectively evaluated. The final diagnosis was chronic pancreatitis (CP) in 20 patients [6 with non-mass-forming pancreatitis, 7 with mass-forming pancreatitis (MFP), and 7 with autoimmune pancreatitis (AIP)], pancreatic cancer (PC) in 72, pancreatic neuroendocrine tumor (PNET) in 9, and normal pancreas in 8. The tissue elasticity distribution calculation was performed in real time, and the results were represented in color in fundamental B-mode imaging. In addition, we performed quantification using the SR (non-mass area/mass area).

RESULTS

Elastography for all PC patients showed intense blue coloration, indicating malignant lesions. In contrast, MFP presented with a mixed coloration pattern of green, yellow, and low-intensity blue. Normal controls showed an even distribution of green to red. The mean SR was 23.66 ± 12.65 for MFP and 39.08 ± 20.54 for PC (P < 0.05).

CONCLUSIONS

Endoscopic ultrasound elastography is a promising diagnostic tool for defining the tissue characteristics of pancreatic masses. In addition, semiquantitative analysis of elasticity using the SR may allow the differentiation of MFP from PC.

In-vitro Strain and Modulus Measurements in Porcine Cervical Lymph Nodes

Cervical lymph nodes are common sites of metastatic involvement in head and neck cancers. These lymph nodes are superficially located and palpation is a common practice for assessing nodal hardness and staging cancer which is, however, too subjective and with limited accuracy. In this study, the mechanical properties of pig lymph node tissues were investigated using ultrasound elastography and indentation test. Lymph nodes were excised from fresh pork pieces and embedded in an agar-gelatin phantom for strain imaging by elastography. A strain ratio reflecting the strain contrast of lymph node over agar-gelatin phantom was used to assess the elasticity of the lymph node. A cutting device was then custom-designed to slice the phantom into uniform slices for indentation test. The measurements revealed that there were significant differences in both the strain ratio and Young’s modulus between the peripheral and middle regions of the lymph nodes (both p < 0.05); however, the results appeared contradictory. Correlation between the results of the two measurements (modulus ratio vs. inversed strain ratio) showed their association was moderate for both the peripheral and middle regions (R² = 0.437 and 0.424 respectively). As the tests were only performed on normal lymph nodes, comparison in stiffness between healthy and abnormal lymph nodes could not be made. Future studies should be conducted to quantify the stiffness change in abnormal lymph nodes.

An improved quantitative measurement for thyroid cancer detection based on elastography

OBJECTIVE

To evaluate color thyroid elastograms quantitatively and objectively.

MATERIALS AND METHODS

125 cases (56 malignant and 69 benign) were collected with the HITACHI Vision 900 system (Hitachi Medical System, Tokyo, Japan) and a liner-array-transducer of 6-13MHz. Standard of reference was cytology (FNA-fine needle aspiration) or histology (core biopsy). The original color thyroid elastograms were transferred from red, green, blue (RGB) color space to hue, saturation, value (HSV) color space. Then, hard area ratio was defined. Finally, a SVM classifier was used to classify thyroid nodules into benign and malignant. The relation between the performance and hard threshold was fully investigated and studied.

RESULTS

The classification accuracy changed with the hard threshold, and reached maximum (95.2%) at some values (from 144 to 152). It was higher than strain ratio (87.2%) and color score (83.2%). It was also higher than the one of our previous study (93.6%).

CONCLUSION

The hard area ratio is an important feature of elastogram, and appropriately selected hard threshold can improve classification accuracy.

State-of-the-art imaging techniques in endoscopic ultrasound

Endoscopic ultrasound (EUS) has recently evolved through technological improvement of equipment, with a major clinical impact in digestive and mediastinal diseases. State-of-the-art EUS equipment now includes real-time sono-elastography, which might be useful for a better characterization of lesions and increased accuracy of differential diagnosis (for e.g. lymph nodes or focal pancreatic lesions). Contrast-enhanced EUS imaging is also available, and is already being used for the differential diagnosis of focal pancreatic masses. The recent development of low mechanical index contrast harmonic EUS imaging offers hope for improved diagnosis, staging and monitoring of anti-angiogenic treatment. Tridimensional EUS (3D-EUS) techniques can be applied to enhance the spatial understanding of EUS anatomy, especially for improved staging of tumors, obtained through a better assessment of the relationship with major surrounding vessels. Despite the progress gained through all these imaging techniques, they cannot replace cytological or histological diagnosis. However, real-time optical histological diagnosis can be achieved through the use of single-fiber confocal laser endomicroscopy techniques placed under real-time EUS-guidance through a 22G needle. Last, but not least, EUS-assisted natural orifice transluminal endoscopic surgery (NOTES) procedures offer a whole new area of imaging applications, used either for combination of NOTES peritoneoscopy and intraperitoneal EUS, but also for access of retroperitoneal organs through posterior EUS guidance.

Differential diagnosis of pancreatic cancer from normal tissue with digital imaging processing and pattern recognition based on a support vector machine of EUS images

BACKGROUND

EUS can detect morphologic abnormalities of pancreatic cancer with high sensitivity but with limited specificity.

OBJECTIVE

To develop a classification model for differential diagnosis of pancreatic cancer by using a digital imaging processing (DIP) technique to analyze EUS images of the pancreas.

DESIGN

A retrospective, controlled, single-center design was used.

SETTING

The study took place at the Second Military Medical University, Shanghai, China.

PATIENTS

There were 153 pancreatic cancer and 63 noncancer patients in this study.

INTERVENTION

All patients underwent EUS-guided FNA and pathologic analysis.

MAIN OUTCOME MEASUREMENTS

EUS images were obtained and correlated with cytologic findings after FNA. Texture features were extracted from the region of interest, and multifractal dimension vectors were introduced in the feature selection to the frame of the M-band wavelet transform. The sequential forward selection process was used for a better combination of features. By using the area under the receiver operating characteristic curve and other texture features based on separability criteria, a predictive model was built, trained, and validated according to the support vector machine theory.

RESULTS

From 67 frequently used texture features, 20 better features were selected, resulting in a classification accuracy of 99.07% after being added to 9 other features. A predictive model was then built and trained. After 50 random tests, the average accuracy, sensitivity, specificity, positive predictive value, and negative predictive value for the diagnosis of pancreatic cancer were 97.98 ± 1.23%, 94.32 ± 0.03%, 99.45 ± 0.01%, 98.65 ± 0.02%, and 97.77 ± 0.01%, respectively.

LIMITATIONS

The limitations of this study include the small sample size and that the support vector machine was not performed in real time.

CONCLUSION

The classification of EUS images for differentiating pancreatic cancer from normal tissue by DIP is quite useful. Further refinements of such a model could increase the accuracy of EUS diagnosis of tumors.

Real-time sono-elastography in the diagnosis of diffuse liver diseases

AIM: To analyze whether computer-enhanced dynamic analysis of elastography movies is able to better characterize and differentiate between different degrees of liver fibrosis.

METHODS: The study design was prospective. A total of 132 consecutive patients with chronic liver diseases and healthy volunteers were examined by transabdominal ultrasound elastography. All examinations were done by two doctors.

RESULTS: Due to the limitations of the method, we obtained high-quality elastography information in only 73.48% of the patients. The κ-means clustering method was applied to assess the inter-observer diagnosis variability, which showed good variability values in accordance with the experience of ultrasound examination of every observer. Cohen’s κ test indicated a moderate agreement between the study observers (κ = 0.4728). Furthermore, we compared the way the two observers clustered the patients, using the test for comparing two proportions (t value, two-sided test). There was no statistically significant difference between the two physicians, regardless of the patients’ real status.

CONCLUSION: Transabdominal real-time elastography is certainly a very useful method in depicting liver hardness, although it is incompletely tested in large multicenter studies.

Real-time ultrasound elastographic imaging of ocular and periocular tissues: a feasibility study

BACKGROUND AND OBJECTIVE

This study examines the value of ultrasound elastography for the examination of ocular and periocular structures.

SUBJECTS AND METHODS

Five patients, aged 22 to 75 years, who each had one blind eye were included. Patients underwent ultrasound elastography of their blind eye and periocular tissues using a 7-13 MHz probe. Strain grayscale and color-coded elastographic maps were recorded. In the former, a quantitative assessment of signal intensity (corresponding to elastic properties) for specific anatomical structures was performed.

RESULTS

Anterior vitreous displayed intermediate elasticity, whereas posterior vitreous displayed low elasticity. Medial and lateral rectus muscle elasticity was higher in primary position than in adduction or abduction.

CONCLUSION

The pattern of elastic imaging in the vitreous cavity could be attributed to posterior vitreous detachment, whereas that of medial and lateral rectus muscles may be related to the level of muscle fiber strain.

Staging role of EUS

Type of lymphoma and stage of disease are the two decisive prognostic factors and therapeutic determinants. For the locoregional staging, i.e. assessment of the gastric wall infiltration and perigastric lymphonodular involvement, endoscopic ultrasound (EUS) is highly useful. EUS has, therefore, to be integrated into the standard staging procedure of gastric lymphoma, although its impact on initial treatment decisions might be limited in the individual case. A benefit from the use of miniechoendoscopes, EUS elastography and EUS-guided biopsies has not yet been proven in gastric lymphoma. EUS also confers an important prognostic value regarding treatment responses to Helicobacter pylori eradication. On the contrary, EUS cannot be recommended as a regular part of follow-up investigations considering its limited value in predicting the response of the lymphoma to radiation or chemotherapy.

In vivo characterization of pancreatic and lymph node tissue by using EUS spectrum analysis: a validation study

BACKGROUND

Quantitative spectral analysis of the radiofrequency (RF) signals that underlie grayscale EUS images can be used to provide additional, objective information about tissue state.

OBJECTIVE

Our purpose was to validate RF spectral analysis as a method to distinguish between (1) benign and malignant lymph nodes and (2) normal pancreas, chronic pancreatitis, and pancreatic cancer.

DESIGN AND SETTING

A prospective validation study of eligible patients was conducted to compare with pilot study RF data.

PATIENTS

Forty-three patients underwent EUS of the esophagus, stomach, pancreas, and surrounding intra-abdominal and mediastinal lymph nodes (19 from a previous pilot study and 24 additional patients).

MAIN OUTCOME MEASUREMENTS

Midband fit, slope, intercept, and correlation coefficient from a linear regression of the calibrated RF power spectra were determined.

RESULTS

Discriminant analysis of mean pilot-study parameters was then performed to classify validation-study parameters. For benign versus malignant lymph nodes, midband fit and intercept (both with t test P < .058) provided classification with 67% accuracy and area under the receiver operating curve (AUC) of 0.86. For diseased versus normal pancreas, midband fit and correlation coefficient (both with analysis of variance P < .001) provided 93% accuracy and an AUC of 0.98. For pancreatic cancer versus chronic pancreatitis, the same parameters provided 77% accuracy and an AUC of 0.89. Results improved further when classification was performed with all data.

LIMITATIONS

Moderate sample size and spatial averaging inherent to the technique.

CONCLUSIONS

This study confirms that mean spectral parameters provide a noninvasive method to quantitatively discriminate benign and malignant lymph nodes as well as normal and diseased pancreas.

The impact of EUS in primary gastric lymphoma

Gastric lymphoma is the most frequent site of gastrointestinal lymphoma and is accessible for endosonographic evaluation. Most primary gastric lymphomas are classified as mucosa-associated lymphoid tissue (MALT)-type lymphomas that develop in the course of chronic Helicobacter pylori infection. Endoscopic ultrasonography (EUS) is regarded to be the most accurate method for the local staging of gastric lymphoma, although scientific evidence is limited. In stage uEI1 low-grade lymphoma, EUS is able to predict a high chance for cure by H. pylori eradication. The significance of EUS elastography or EUS-guided fine-needle aspiration biopsy to diagnose nodal involvement has not been investigated in prospective series yet. Since high-grade lymphoma is always treated as a systemic disease, the impact of endosonographic staging is lower than in low-grade lymphoma. After treatment of primary gastric lymphoma, EUS produces conflicting results that are not as accurate as endoscopy with biopsy. Therefore, EUS is not mandatory during follow-up.

The role of sonoelastography in the differential diagnosis of neck nodules

Sonoelastography is an imaging technique that provides information on tissue elasticity. Its use as a diagnostic procedure is based on the premise that pathological processes like cancer alter the physical characteristics of the involved tissue. Ultrasonographic studies of the neck can reveal the nonpalpable thyroid nodules, but the nature of these lesions generally has to be established on the basis of FNAB findings. In our hands, sonoelastography displayed a diagnostic accuracy of 86.2% in identifying thyroid nodule malignancy, with positive and negative predictive values (PPV and NPV) of 64% and 94.5%, respectively. In the study of cervical lymph nodes, the results were less impressive (sensitivity 75%, specificity 80%, accuracy 77%, PPV 80%, NPV 70%), but the information obtained with this technique can in our opinion be a useful adjunct to sonographic findings. Indeed, in 5 lymph nodes with sonographic features consistent with malignancy, sonoelastography revealed diffuse elasticity that was indicative of benign disease, which was confirmed by pathological studies. Other nodular lesions of the neck can also be evaluated with sonoelastography, including enlarged parotid glands, but the data in the literature are too limited to allow hypotheses on the role of this imaging modality in this field. Sonoelastography is rapid and simple to perform, and it appears to be a potentially useful tool for the differential diagnosis of neck nodules. This is particularly true of thyroid nodules. Our experience with these lesions indicates that diffuse elasticity is strongly correlated with benign disease. If this finding is confirmed in larger studies, sonoelastography might be used to identify thyroid nodules that do not require immediate biopsy.

Esophageal cancer: ultrasonography

Esophageal malignancy is a major source of morbidity and mortality, despite the recently increased attention to screening and early detection. Prognosis for esophageal cancer remains grim, with advanced tumor stage and lymph node metastases conferring even graver outcomes. Several studies have demonstrated that the addition of preoperative neoadjuvant chemoradiotherapy may improve survival in patients with locally advanced tumor (T3) disease or local lymph node metastases. It is here that endoscopic ultrasonography finds its niche in the precise staging of these tumors and the subsequent use of stage-dependent treatment protocols.

Neural network analysis of dynamic sequences of EUS elastography used for the differential diagnosis of chronic pancreatitis and pancreatic cancer

BACKGROUND

EUS elastography is a newly developed imaging procedure that characterizes the differences of hardness and strain between diseased and normal tissue.

OBJECTIVE

To assess the accuracy of real-time EUS elastography in pancreatic lesions.

DESIGN

Cross-sectional feasibility study.

PATIENTS

The study group included, in total, 68 patients with normal pancreas (N = 22), chronic pancreatitis (N = 11), pancreatic adenocarcinoma (N = 32), and pancreatic neuroendocrine tumors (N = 3). A subgroup analysis of 43 cases with focal pancreatic masses was also performed.

INTERVENTIONS

A postprocessing software analysis was used to examine the EUS elastography movies by calculating hue histograms of each individual image, data that were further subjected to an extended neural network analysis to differentiate benign from malignant patterns.

MAIN OUTCOME MEASUREMENTS

To differentiate normal pancreas, chronic pancreatitis, pancreatic cancer, and neuroendocrine tumors.

RESULTS

Based on a cutoff of 175 for the mean hue histogram values recorded on the region of interest, the sensitivity, specificity, and accuracy of differentiation of benign and malignant masses were 91.4%, 87.9%, and 89.7%, respectively. The positive and negative predictive values were 88.9% and 90.6%, respectively. Multilayer perceptron neural networks with both one and two hidden layers of neurons (3-layer perceptron and 4-layer perceptron) were trained to learn how to classify cases as benign or malignant, and yielded an excellent testing performance of 95% on average, together with a high training performance that equaled 97% on average.

LIMITATION

A lack of the surgical standard in all cases.

CONCLUSIONS

EUS elastography is a promising method that allows characterization and differentiation of normal pancreas, chronic pancreatitis, and pancreatic cancer. The currently developed methodology, based on artificial neural network processing of EUS elastography digitalized movies, enabled an optimal prediction of the types of pancreatic lesions. Future multicentric, randomized studies with adequate power will have to establish the clinical impact of this procedure for the differential diagnosis of focal pancreatic masses.