RESUMO
OBJECTIVE: Objective measurement of pre-operative severity is important to optimize evidence-based practices given that the wide spectrum of presentation likely influences outcomes. The purpose of this study was to determine the correlation of objective measures of form with a subjective standard of cleft severity. DESIGN: 3D images were ranked according to severity of nasal deformity by 7 cleft surgeons so that the mean rank could be used as the severity standard. PATIENTS: 45 patients with unilateral cleft lip and 5 normal control subjects. INTERVENTIONS: Each image was assessed using traditional anthropometric analysis, 3D landmark displacements, and shape-based analysis to produce 81 indices for each subject. MAIN OUTCOME: The correlation of objective measurements with the clinical severity standard. RESULTS: Lateral deviation of subnasale from midline was the best predictor of severity (0.86). Other strongly-correlated anthropometric measurements included columellar angle, nostril width ratio, and lateral lip height ratio (0.72, 0.80, 0.79). Almost all shape-based measurements had tight correlation with the severity standard, however, dorsum deviation and point difference nasolabial symmetry were the most predictive (0.84, 0.82). CONCLUSIONS: Quantitative measures of severity transcend cleft type and can be used to grade clinical severity. Lateral deviation of subnasale was the best measure of severity and may be used as a surrogate of uncoupled premaxillary growth; it should be recorded as an index of pre-operative severity with every cleft lip repair. The correlation of other measures evaluated clarify treatment priorities and could potentially be used to grade outcomes.
RESUMO
BACKGROUND: Modifications of the Pi craniectomy technique are meant to address the occipital bullet deformity of sagittal synostosis but it is not clear if they result in persistent improvement. Our purpose was to use morphometric analysis to determine if a low occipital osteotomy with verticalization results in improved occipital shape after a modified pi procedure two years after surgery. METHOD: We performed a retrospective cohort study comparing modified Pi technique with and without a low occipital osteotomy with verticalization immediately and two years after surgery relative to age-matched normal controls. We used anthropometric measures and population-level anatomical templates using multivariate template construction script from Advanced Normalization Tools for comparison between groups. A subgroup analysis was performed for severe occipital bullet deformity at presentation. RESULTS: We observed stable improvement in the angle of the inferior occiput with the occipital remodeling modification that persisted two years after surgery. This improvement was seen in the entire cohort and was greater in the severe sub-group analysis. Complications and blood transfusion volumes were not different between the two techniques. The LOOV group demonstrated improved posterior vertical height and cephalic index immediately after surgery, but these did not persist two years later. CONCLUSION: Occipital remodeling improves the bullet deformity but does not affect posterior vertical height two years after surgery. We recommend direct inferior occipital remodeling when using the Pi technique for young patients with acute occipital incline angles and occipital constriction.
RESUMO
BACKGROUND: Open middle and posterior cranial vault expansion (OPVE) or endoscopic (ES) strip craniectomy are two surgical techniques for normalization of head shape in isolated sagittal synostosis. This study aims to compare two-year cranial morphometrics after these two approaches. METHODS: We performed morphometric analysis on preoperative (t0), immediately post-operative (t1) and 2-year (t2) postoperative CT scans of patients who underwent OPVE or ES prior to 4 months of age. Perioperative data and morphometrics were compared between the two groups and age-matched controls. RESULTS: Nineteen patients were included in the ES cohort, 19 age-matched patients in the OPVE cohort, and 57 as controls. Median surgery time and blood transfusion volume were less for the ES approach (118 min; 0cc) compared to OPVE (204 min; 250cc). Anthropometric measurements after OPVE were closer normal controls at t1 compared to ES, but the skull shapes were comparable at t2. In the mid-sagittal plane, anterior vault was higher after OPVE at t2 compared to both ES and controls, but the posterior length was shorter and closer to controls than the ES cohort. Cranial volumes were like controls for both cohorts at t2. There was no difference in complication rate. CONCLUSIONS: Both OPVE and ES techniques result in normalization of cranial shape in patients with isolated sagittal synostosis after two years with minimal morphometric differences. Family decision-making between the two approaches should be based on age at presentation, avoidance of blood transfusion, scar pattern, and availability of helmet molding and not on expected outcome. LEVEL OF EVIDENCE: III.
RESUMO
BACKGROUND: The purpose of this study was to quantify change in cranial morphology in patients with nonsyndromic unilateral lambdoid craniosynostosis (ULC) from presentation (t0), after open posterior switch-cranioplasty (t1), and at 2-year follow-up (t2). METHODS: Volumetric, linear, and angular analysis were performed on computed tomographic scans at the three time points and against normal control subjects. Significance was set at P < 0.05. RESULTS: Twenty-two patients were included. ULC cranial vault asymmetry index was higher than in control subjects before surgery (6.22 ± 3.55) but decreased after surgery (3.00 ± 2.53) to become comparable with the normal asymmetry range present in the controls. After surgery, both diagonals increased, but more on the fused side. In the 2 years after surgery, both diagonals in patients with ULC grew proportionately, but the fused diagonal remained slightly shorter than the patent side. Total cranial volume was higher in patients with ULC than in control subjects after surgery but became comparable at t2. Cranial base angulation improved by t2 but did not approach normal, and ear position remained unchanged. The facial twist was higher than in controls at t0 and t1 but was comparable at t2. Coronal asymmetry improved with surgery but remained undercorrected at t2, with the greatest residual asymmetry at opisthion. CONCLUSIONS: Open-switch cranioplasty normalizes cranial vault asymmetry index by increasing the fused cranial diagonal more than the patent side and is stable at 2 years. Skull base twist does not normalize, but facial twist approaches normal. Technique improvement should focus on residual coronal asymmetry present at opisthion. CLINICAL QUESTION/LEVEL OF EVIDENCE: Therapeutic, IV.
Assuntos
Craniossinostoses , Crânio , Humanos , Lactente , Crânio/diagnóstico por imagem , Crânio/cirurgia , Craniossinostoses/diagnóstico por imagem , Craniossinostoses/cirurgia , Base do Crânio/cirurgia , Face/cirurgia , Tomografia Computadorizada por Raios X/métodos , Suturas Cranianas/diagnóstico por imagem , Suturas Cranianas/cirurgiaRESUMO
BACKGROUND: This study aimed to quantify the change in three-dimensional skull morphometrics for patients with sagittal synostosis at presentation, after surgery, and at 2-year follow-up. METHODS: Computed tomography scans from 91 patients with isolated SS were age-, sex-, and race-matched with those from 273 controls. The authors performed vector analysis with linear regressions to model the effect of open middle and posterior cranial vault remodeling on cranial shape and growth. RESULTS: Anterior cranial volume, bossing angle, and frontal shape were not changed by surgery but normalized without surgical intervention by 2 years. Biparietal narrowing and middle cranial volume were corrected after surgery and maintained at 2 years. Occipital protuberance was improved after surgery and normalized at 2 years. Posterior cranial volume was decreased by occipital remodeling and remained slightly lower than control volumes at 2 years, whereas middle vault volume was larger than in controls. Residual deformities that persisted at 2 years were decreased superolateral width at the level of opisthion and increased anterosuperior height (vertex bulge). Linear models suggested older age at surgery resulted in more scaphocephaly and enlarged posterior cranial vault volumes at 2 years but did not affect other volume outcomes. Preoperative severity was the variable most predictive of 2-year morphometrics. CONCLUSIONS: Initial severity of sagittal synostosis deformity was the best predictor of 2-year morphometric outcomes. Upper posterior cranial width decreases with time after surgery and an anterior vertex bulge can persist after open surgery, but frontal dysmorphology self-corrects without surgical intervention. CLINICAL QUESTION/LEVEL OF EVIDENCE: Therapeutic, IV.
Assuntos
Craniossinostoses , Procedimentos de Cirurgia Plástica , Humanos , Lactente , Crânio/diagnóstico por imagem , Crânio/cirurgia , Craniossinostoses/diagnóstico por imagem , Craniossinostoses/cirurgia , Tomografia Computadorizada por Raios X/métodos , Cabeça/cirurgia , Estudos RetrospectivosRESUMO
BACKGROUND: The timing of posterior cranial expansion for the management of intracranial pressure can be "staged" by age and dysmorphology or "expectant" by pressure monitoring. The authors report shared outcome measures from one center performing posterior vault remodeling (PCVR) or distraction (PVDO) following a staged approach and another performing spring-assisted expansion (SAPVE) following an expectant protocol. METHODS: Apert or Crouzon syndrome patients who underwent posterior expansion younger than 2 years were included. Perioperative outcomes and subsequent cranial operations were recorded up to last follow-up and intracranial volume changes measured and adjusted using growth curves. RESULTS: Thirty-eight patients were included. Following the expectant protocol, Apert patients underwent SAPVE at a younger age (8 months) than Crouzon patients (16 months). The initial surgery time was shorter but total operative time, including device removal, was longer for PVDO (3 hours 52 minutes) and SAPVE (4 hours 34 minutes) than for PCVR (3 hours 24 minutes). Growth-adjusted volume increase was significant and comparable. Fourteen percent of PCVR, 33% of PVDO, and 11% of SAPVE cases had complications, but without long-term deficits. Following the staged approach, 5% underwent only PVDO, 85% had a staged posterior followed by anterior surgery, and 10% required a third expansion. Following the expectant approach, 42% of patients had only posterior expansion at last follow-up, 32% had a secondary cranial surgery, and 26% had a third cranial expansion. CONCLUSION: Two approaches involving posterior vault expansion in young syndromic patients using three techniques resulted in comparable early volume expansion and complication profiles. CLINICAL QUESTION/LEVEL OF EVIDENCE: Therapeutic, III.
Assuntos
Disostose Craniofacial , Craniossinostoses , Osteogênese por Distração , Humanos , Lactente , Craniossinostoses/cirurgia , Crânio/cirurgia , Disostose Craniofacial/cirurgia , Avaliação de Resultados em Cuidados de Saúde , Osteogênese por Distração/métodosRESUMO
BACKGROUND: Balance is achieved through opposing interactions. Objective three-dimensional assessment of changes during surgical treatment of the unilateral cleft lip and nasal deformity are limited, and false assumptions may prevent optimal management. METHODS: The authors performed anthropometric analysis on the immediate preoperative and postoperative images (captured under anesthetic) of patients undergoing primary repair ( n = 36). Changes in dimensions and measures of balance were assessed ( P < 0.05). RESULTS: Angles and ratios that reflect cleft to noncleft side balance normalized, although alterations occurred in opposing ways. Centralization of the columella narrowed the cleft nasal base and widened the noncleft nasal base. As the cleft columellar height elongated, the noncleft columellar height shortened. With these changes and correction of cleft alar base retrusion, the cleft alar dome was raised. The cleft and noncleft lateral lip heights and widths elongated. Meanwhile, the Cupid's bow broadened as the commissures were drawn closer together. Whereas the cleft philtral height lengthened, the noncleft philtral height shortened. Reduction in noncleft philtral height averaged 20% but varied with measures of preoperative severity including columellar angle ( R = 0.67), the difference in philtral heights ( R = 0.65), and lateral deviation of the subnasale ( R = 0.74). CONCLUSIONS: Tissue does not need to be added to "lengthen" the columella, the noncleft philtral height shortening can be estimated, and the contours of anatomic subunits change with surgery on both cleft and noncleft sides. It is inadequate to focus on correction of the cleft side alone without considering corresponding noncleft side changes. Achieving balance through opposing alterations should be the principal goal of treatment. CLINICAL QUESTION/LEVEL OF EVIDENCE: Therapeutic, IV.
Assuntos
Fenda Labial , Doenças Nasais , Procedimentos de Cirurgia Plástica , Humanos , Fenda Labial/cirurgia , Nariz/cirurgia , Septo Nasal/cirurgia , Lábio/cirurgia , Doenças Nasais/cirurgia , Resultado do TratamentoRESUMO
PURPOSE: PIK3CA-related overgrowth spectrum (PROS) conditions of the head and neck are treatment challenges. Traditionally, these conditions require multiple invasive interventions, with incomplete malformation removal, disfigurement, and possible dysfunction. Use of the PI3K inhibitor alpelisib, previously shown to be effective in PROS, has not been reported in PIK3CA-associated head and neck lymphatic malformations (HNLMs) or facial infiltrating lipomatosis (FIL). We describe prospective treatment of 5 children with PIK3CA-associated HNLMs or head and neck FIL with alpelisib monotherapy. METHODS: A total of 5 children with PIK3CA-associated HNLMs (n = 4) or FIL (n = 1) received alpelisib monotherapy (aged 2-12 years). Treatment response was determined by parental report, clinical evaluation, diary/questionnaire, and standardized clinical photography, measuring facial volume through 3-dimensional photos and magnetic resonance imaging. RESULTS: All participants had reduction in the size of lesion, and all had improvement or resolution of malformation inflammation/pain/bleeding. Common invasive therapy was avoided (ie, tracheotomy). After 6 or more months of alpelisib therapy, facial volume was reduced (range 1%-20%) and magnetic resonance imaging anomaly volume (range 0%-23%) were reduced, and there was improvement in swallowing, upper airway patency, and speech clarity. CONCLUSION: Individuals with head and neck PROS treated with alpelisib had decreased malformation size and locoregional overgrowth, improved function and symptoms, and fewer invasive procedures.
Assuntos
Fosfatidilinositol 3-Quinases , Tiazóis , Criança , Humanos , Fosfatidilinositol 3-Quinases/genética , Mutação , Classe I de Fosfatidilinositol 3-Quinases/genética , Tiazóis/uso terapêuticoRESUMO
BACKGROUND: It is important to determine whether sagittal synostosis-associated scaphocephaly is static in the presurgical period, or whether there are morphologic differences with time to include in surgical decision-making. The authors' purpose was to perform cross-sectional analysis of cranial morphology before any surgical intervention in children with sagittal synostosis younger than 9 months compared to matched controls. METHODS: The authors performed morphometric analysis on computed tomographic scans from 111 untreated isolated sagittal synostosis patients younger than 9 months and 37 age-matched normal controls. The authors divided the patients into three age groups and performed statistical comparison between sagittal synostosis and controls for each group. RESULTS: Sagittal synostosis cephalic indices were stable and lower in patients than in controls across groups. Total cranial volume was equivalent, but sagittal synostosis patients had a greater posterior volume than controls at all ages and a smaller middle fossa volume at older ages. Pterional width was greater in sagittal synostosis patients than in controls for each age group. Frontal bossing vectors were most severe in the youngest age groups and least in the older group. Occipital protuberance was consistent across the age groups. CONCLUSIONS: Upper parietal narrowing and occipital protuberance were the consistent deformities across age groups, with the most parietal constriction seen in older patients. Frontal bossing was not consistent and was more severe in the younger patients. The authors did not detect significant pterional constriction, and the appearance of constriction is relative to adjacent morphology and not absolute. CLINICAL QUESTION/LEVEL OF EVIDENCE: Risk, II.
Assuntos
Craniossinostoses , Idoso , Criança , Craniossinostoses/diagnóstico por imagem , Craniossinostoses/cirurgia , Estudos Transversais , Ossos Faciais , Humanos , Lactente , Estudos Retrospectivos , Crânio/cirurgia , Tomografia Computadorizada por Raios X/métodosRESUMO
PURPOSE: A recent modification of the sagittal split osteotomy (SSO) utilizes a medial cut placed at the level of the mandibular occlusal plane (low) and terminating anterior to the retrolingular fossa (short). The purpose of this work was to evaluate the position of the inferior alveolar nerve (IAN) relative to the medial mandibular ramus cortex in patients undergoing SSO utilizing a low medial cut. METHODS: This was a retrospective, cross-sectional evaluation of patients evaluated at a tertiary-care center for facial skeletal deformities who were candidates for mandibular orthognathic surgery. Patients were included as study subjects if they had medical-grade computed tomography (CT) scans as part of preoperative planning. The measure of interest was the closest distance from the medial cortex to the IAN for a horizontal osteotomy placed at the level of the mandibular occlusal plane, as measured on CT scans using an automated process. Descriptive statistics were computed to identify the proximity of the IAN to the medial cortex as a function of the length of the medial horizontal osteotomy. RESULTS: Forty-seven patients (94 SSO sites) with a mean age of 18.7 ± 3.3 years were included as study subjects. Twenty-six subjects had a primary diagnosis of congenital craniofacial anomaly; 21 subjects had a primary dentofacial deformity. For medial ramus osteotomy lengths of 7.5 mm, 10 mm, 12.5 mm, 15 mm, and 20 mm, the closest distances to the IANs were 9.8 ± 2.5 mm, 8.3 ± 2.5 mm, 6.5 ± 2.4 mm, 4.8 ± 2.1 mm, and 2.9 ± 1.6 mm, respectively. CONCLUSION: When utilizing the low medial cut in the SSO, the IAN is reliably found ≥ 5 mm away from the medial cortex of the ramus when the osteotomy length is < 15 mm.
Assuntos
Mandíbula , Osteotomia Sagital do Ramo Mandibular , Adolescente , Adulto , Estudos Transversais , Humanos , Mandíbula/diagnóstico por imagem , Mandíbula/cirurgia , Nervo Mandibular/diagnóstico por imagem , Estudos Retrospectivos , Adulto JovemRESUMO
BACKGROUND: Although many cleft teams have adopted nasoalveolar molding to improve nasal form, few comparative studies have assessed the postoperative benefits of this treatment. Given that reported outcomes have been contradictory and that treatment involves considerable burden to families, the purpose of this study was to assess objective and subjective changes from nasoalveolar molding at approximately 5 years of age. METHODS: All patients with complete unilateral cleft lip and palate who underwent primary cheiloplasty performed by a single surgeon over a 7-year period were reviewed. Patient results were grouped into nasoalveolar molding or no-nasoalveolar molding. Cleft severity and aesthetic outcomes were assessed by panels of raters who independently ranked subject images at presentation, immediately preoperative (after molding), and at 5-year follow-up. Objective symmetry was measured using standard anthropometric analysis on three-dimensional images. RESULTS: Among 41 patients included, 16 successfully completed nasoalveolar molding. Both groups were similar at presentation; however, the nasoalveolar molding group had improved appearance following molding (p < 0.05). After surgery, at 5 years of age, the nasoalveolar molding group had better rank scores for overall appearance (p < 0.05), cleft nostril height, and cleft medial lip height (p < 0.05). Regression analysis revealed that nasoalveolar molding treatment was the most significant predictor of overall nasal appearance at 5 years, but that treatment team experience and initial severity were also significant predictors (p < 0.05). Qualitative audit following analysis identified favorable and unfavorable features of nasoalveolar molding. CONCLUSION: In children with complete unilateral cleft lip and palate, nasoalveolar molding was associated with better overall nasal aesthetics and improved cleft nostril height and cleft medial lip height at approximately 5 years of age. CLINICAL QUESTION/LEVEL OF EVIDENCE: Therapeutic, III.
Assuntos
Fenda Labial/terapia , Fissura Palatina/terapia , Moldagem Nasoalveolar/estatística & dados numéricos , Nariz/anatomia & histologia , Procedimentos de Cirurgia Plástica/estatística & dados numéricos , Criança , Pré-Escolar , Estética , Feminino , Seguimentos , Humanos , Lactente , Recém-Nascido , Masculino , Nariz/cirurgia , Estudos Retrospectivos , Resultado do TratamentoRESUMO
Anatomic studies have identified that patients with Treacher Collins syndrome and some cases of bilateral craniofacial microsomia are characterized by multilevel airway obstruction as a result of hypoplasia and clockwise rotation of the maxillomandibular complex. Patients often remain tracheostomy-dependent despite multiple airway surgeries. Counterclockwise craniofacial distraction osteogenesis aims to correct the facial skeletal deformity and expand the upper airway volume by rotating the subcranial complex en bloc around the nasofrontal junction. Early results have demonstrated significant increases in the nasopharyngeal and oropharyngeal airway volumes with successful decannulation in a majority of patients who have undergone this operation.
Assuntos
Síndrome de Goldenhar/cirurgia , Disostose Mandibulofacial/cirurgia , Osteogênese por Distração/métodos , Obstrução das Vias Respiratórias/etiologia , Obstrução das Vias Respiratórias/cirurgia , Ossos Faciais/anormalidades , Ossos Faciais/cirurgia , Feminino , Humanos , Masculino , Mandíbula/cirurgia , TraqueostomiaRESUMO
Severe midface hypoplasia is often managed by Le Fort III distraction. Le Fort II distraction with zygomatic repositioning is a modification of the Le Fort III distraction operation aimed to correct abnormal facial ratios of patients with greater central than lateral midface deficiency. The operation starts with Le Fort III osteotomies and is followed by separation and fixation of bilateral zygomas. The central nasomaxillary Le Fort II segment is then distracted to achieve independent movements of the central and lateral midface. The Le Fort II zygomatic repositioning operation has become our procedure of choice for patients with Apert facial dysmorphology.
Assuntos
Acrocefalossindactilia/cirurgia , Osteotomia de Le Fort/métodos , Zigoma/cirurgia , Ossos Faciais/anormalidades , Ossos Faciais/cirurgia , Feminino , Humanos , Masculino , Maxila/cirurgia , Osteogênese por Distração/métodosRESUMO
BACKGROUND: Current descriptions of the unilateral cleft lip and nasal deformity (uCLND) are based upon limited sample sets and subjective observations. While those descriptions are inconsistent and contradictory, theoretical models, including Hogan's "tilted tripod" and Fisher's "nasal arch forms", have never been tested. Given that favorable outcomes of treatment remain elusive, detailed study of the deformity is critical in devising better treatments. The purpose of this study was to develop a data-driven three-dimensional (3D) model of uCLND that spans the spectrum of presentation and involves a pervasive underlying mechanism. METHODS: We studied 3D images of 100 infants with unrepaired cleft lip at 6 months of age. Objective assessment included the landmark positions, anthropometric dimensions, and shape-based measures. Cleft severity was stratified by the lateral displacement of subnasale, so that a model could be developed using linear regression. RESULTS: With progressive deviation of subnasale, the non-cleft alar base moved lateral, the cleft alar base was left posterior, and the nasal dorsum followed the caudal septum (deviating towards the non-cleft side). The "twist" resulted in opposing cleft alar dome collapse, non-cleft alar ring constriction (the non-cleft nasal sill narrowed, lateral genu rose, and alar-cheek junction became more acute), and displacement of the philtrum from midline. CONCLUSIONS: Our study not only supports theoretical models of uCLND but also clarifies vectors of change and reveals significant non-cleft side alterations. On the basis of our findings, the objectives of treatment should involve centralization of the columella and philtrum and rebalancing the nose by untwisting the orthogonal displacement of the alar bases.
Assuntos
Fenda Labial/diagnóstico por imagem , Modelos Anatômicos , Nariz/diagnóstico por imagem , Pontos de Referência Anatômicos , Antropometria , Fenda Labial/patologia , Fenda Labial/cirurgia , Feminino , Humanos , Imageamento Tridimensional , Lactente , Lábio/diagnóstico por imagem , Lábio/patologia , Masculino , Modelos Teóricos , Nariz/anormalidades , Nariz/patologiaRESUMO
PURPOSE: Subcranial midface distraction is used to treat central midface deficiency in syndromic synostosis. Our aim was to determine which maxillary movements were associated with improvement in measures of obstructive sleep apnea. METHODS: This was a retrospective cohort study that reviewed patients with syndromic midface retrusion and documented sleep apnea who underwent subcranial midface distraction via either Le Fort 3 osteotomy or Le Fort 2 osteotomy with zygomatic repositioning. The predictor variables measured on cephalograms were the magnitude and direction of midface and mandibular movements. The primary outcome was the change in the apnea hypopnea index (AHI) from polysomnography before and after surgery. The secondary outcomes were volumes of upper airway containing bone spaces calculated from computed tomography scans. Data analysis included linear regression to estimate the effect of distraction vectors on bone space volumes and AHI changes. RESULTS: We included 18 patients primarily with Crouzon or Apert syndrome. The magnitude of distraction in a horizontal direction was the most significant factor for AHI improvement and primarily expanded the nasopharyngeal space, but with a smaller impact on the oral cavity space. Clockwise palate rotation was most influenced by a downward direction of distraction, with 24° below horizontal creating a neutral advancement. The greater the magnitude of advancement, the more likely a counterclockwise rotation was observed. CONCLUSIONS: Horizontal magnitude of advancement had the greatest impact on AHI improvement. Vertical lengthening and closure of anterior open bite deformities can be done without compromising airway results as long as total advancement is not compromised. Palate rotation is best controlled by a downward distraction vector, but counterclockwise rotation increases with greater advancement.
Assuntos
Osteogênese por Distração , Cefalometria , Humanos , Osteotomia de Le Fort , Estudos Retrospectivos , RotaçãoRESUMO
In this study, we propose the Ductal Instance-Oriented Pipeline (DIOP) that contains a duct-level instance segmentation model, a tissue-level semantic segmentation model, and three-levels of features for diagnostic classification. Based on recent advancements in instance segmentation and the Mask RCNN model, our duct-level segmenter tries to identify each ductal individual inside a microscopic image; then, it extracts tissue-level information from the identified ductal instances. Leveraging three levels of information obtained from these ductal instances and also the histopathology image, the proposed DIOP outperforms previous approaches (both feature-based and CNN-based) in all diagnostic tasks; for the four-way classification task, the DIOP achieves comparable performance to general pathologists in this unique dataset. The proposed DIOP only takes a few seconds to run in the inference time, which could be used interactively on most modern computers. More clinical explorations are needed to study the robustness and generalizability of this system in the future.
RESUMO
BACKGROUND: Pathologists analyze biopsy material at both the cellular and structural level to determine diagnosis and cancer stage. Mitotic figures are surrogate biomarkers of cellular proliferation that can provide prognostic information; thus, their precise detection is an important factor for clinical care. Convolutional Neural Networks (CNNs) have shown remarkable performance on several recognition tasks. Utilizing CNNs for mitosis classification may aid pathologists to improve the detection accuracy. METHODS: We studied two state-of-the-art CNN-based models, ESPNet and DenseNet, for mitosis classification on six whole slide images of skin biopsies and compared their quantitative performance in terms of sensitivity, specificity, and F-score. We used raw RGB images of mitosis and non-mitosis samples with their corresponding labels as training input. In order to compare with other work, we studied the performance of these classifiers and two other architectures, ResNet and ShuffleNet, on the publicly available MITOS breast biopsy dataset and compared the performance of all four in terms of precision, recall, and F-score (which are standard for this data set), architecture, training time and inference time. RESULTS: The ESPNet and DenseNet results on our primary melanoma dataset had a sensitivity of 0.976 and 0.968, and a specificity of 0.987 and 0.995, respectively, with F-scores of .968 and .976, respectively. On the MITOS dataset, ESPNet and DenseNet showed a sensitivity of 0.866 and 0.916, and a specificity of 0.973 and 0.980, respectively. The MITOS results using DenseNet had a precision of 0.939, recall of 0.916, and F-score of 0.927. The best published result on MITOS (Saha et al. 2018) reported precision of 0.92, recall of 0.88, and F-score of 0.90. In our architecture comparisons on MITOS, we found that DenseNet beats the others in terms of F-Score (DenseNet 0.927, ESPNet 0.890, ResNet 0.865, ShuffleNet 0.847) and especially Recall (DenseNet 0.916, ESPNet 0.866, ResNet 0.807, ShuffleNet 0.753), while ResNet and ESPNet have much faster inference times (ResNet 6 s, ESPNet 8 s, DenseNet 31 s). ResNet is faster than ESPNet, but ESPNet has a higher F-Score and Recall than ResNet, making it a good compromise solution. CONCLUSION: We studied several state-of-the-art CNNs for detecting mitotic figures in whole slide biopsy images. We evaluated two CNNs on a melanoma cancer dataset and then compared four CNNs on a public breast cancer data set, using the same methodology on both. Our methodology and architecture for mitosis finding in both melanoma and breast cancer whole slide images has been thoroughly tested and is likely to be useful for finding mitoses in any whole slide biopsy images.