RESUMO
The rapid development of artificial intelligence (AI) has gained importance, with many tools already entering our daily lives. The medical field of radiation oncology is also subject to this development, with AI entering all steps of the patient journey. In this review article, we summarize contemporary AI techniques and explore the clinical applications of AI-based automated segmentation models in radiotherapy planning, focusing on delineation of organs at risk (OARs), the gross tumor volume (GTV), and the clinical target volume (CTV). Emphasizing the need for precise and individualized plans, we review various commercial and freeware segmentation tools and also state-of-the-art approaches. Through our own findings and based on the literature, we demonstrate improved efficiency and consistency as well as time savings in different clinical scenarios. Despite challenges in clinical implementation such as domain shifts, the potential benefits for personalized treatment planning are substantial. The integration of mathematical tumor growth models and AI-based tumor detection further enhances the possibilities for refining target volumes. As advancements continue, the prospect of one-stop-shop segmentation and radiotherapy planning represents an exciting frontier in radiotherapy, potentially enabling fast treatment with enhanced precision and individualization.
RESUMO
BACKGROUND: Volume of interest (VOI) segmentation is a crucial step for Radiomics analyses and radiotherapy (RT) treatment planning. Because it can be time-consuming and subject to inter-observer variability, we developed and tested a Deep Learning-based automatic segmentation (DLBAS) algorithm to reproducibly predict the primary gross tumor as VOI for Radiomics analyses in extremity soft tissue sarcomas (STS). METHODS: A DLBAS algorithm was trained on a cohort of 157 patients and externally tested on an independent cohort of 87 patients using contrast-enhanced MRI. Manual tumor delineations by a radiation oncologist served as ground truths (GTs). A benchmark study with 20 cases from the test cohort compared the DLBAS predictions against manual VOI segmentations of two residents (ERs) and clinical delineations of two radiation oncologists (ROs). The ROs rated DLBAS predictions regarding their direct applicability. RESULTS: The DLBAS achieved a median dice similarity coefficient (DSC) of 0.88 against the GTs in the entire test cohort (interquartile range (IQR): 0.11) and a median DSC of 0.89 (IQR 0.07) and 0.82 (IQR 0.10) in comparison to ERs and ROs, respectively. Radiomics feature stability was high with a median intraclass correlation coefficient of 0.97, 0.95 and 0.94 for GTs, ERs, and ROs, respectively. DLBAS predictions were deemed clinically suitable by the two ROs in 35% and 20% of cases, respectively. CONCLUSION: The results demonstrate that the DLBAS algorithm provides reproducible VOI predictions for radiomics feature extraction. Variability remains regarding direct clinical applicability of predictions for RT treatment planning.
Assuntos
Algoritmos , Benchmarking , Aprendizado Profundo , Extremidades , Imageamento por Ressonância Magnética , Sarcoma , Humanos , Sarcoma/diagnóstico por imagem , Sarcoma/radioterapia , Sarcoma/patologia , Imageamento por Ressonância Magnética/métodos , Masculino , Feminino , Extremidades/diagnóstico por imagem , Pessoa de Meia-Idade , Adulto , Idoso , Planejamento da Radioterapia Assistida por Computador/métodos , Neoplasias de Tecidos Moles/diagnóstico por imagem , Neoplasias de Tecidos Moles/radioterapia , Neoplasias de Tecidos Moles/patologia , RadiômicaRESUMO
BACKGROUND: Many automatic approaches to brain tumor segmentation employ multiple magnetic resonance imaging (MRI) sequences. The goal of this project was to compare different combinations of input sequences to determine which MRI sequences are needed for effective automated brain metastasis (BM) segmentation. METHODS: We analyzed preoperative imaging (T1-weighted sequence ± contrast-enhancement (T1/T1-CE), T2-weighted sequence (T2), and T2 fluid-attenuated inversion recovery (T2-FLAIR) sequence) from 339 patients with BMs from seven centers. A baseline 3D U-Net with all four sequences and six U-Nets with plausible sequence combinations (T1-CE, T1, T2-FLAIR, T1-CE + T2-FLAIR, T1-CE + T1 + T2-FLAIR, T1-CE + T1) were trained on 239 patients from two centers and subsequently tested on an external cohort of 100 patients from five centers. RESULTS: The model based on T1-CE alone achieved the best segmentation performance for BM segmentation with a median Dice similarity coefficient (DSC) of 0.96. Models trained without T1-CE performed worse (T1-only: DSC = 0.70 and T2-FLAIR-only: DSC = 0.73). For edema segmentation, models that included both T1-CE and T2-FLAIR performed best (DSC = 0.93), while the remaining four models without simultaneous inclusion of these both sequences reached a median DSC of 0.81-0.89. CONCLUSIONS: A T1-CE-only protocol suffices for the segmentation of BMs. The combination of T1-CE and T2-FLAIR is important for edema segmentation. Missing either T1-CE or T2-FLAIR decreases performance. These findings may improve imaging routines by omitting unnecessary sequences, thus allowing for faster procedures in daily clinical practice while enabling optimal neural network-based target definitions.
RESUMO
BACKGROUND: Stereotactic radiotherapy is a standard treatment option for patients with brain metastases. The planning target volume is based on gross tumor volume (GTV) segmentation. The aim of this work is to develop and validate a neural network for automatic GTV segmentation to accelerate clinical daily routine practice and minimize interobserver variability. METHODS: We analyzed MRIs (T1-weighted sequence ± contrast-enhancement, T2-weighted sequence, and FLAIR sequence) from 348 patients with at least one brain metastasis from different cancer primaries treated in six centers. To generate reference segmentations, all GTVs and the FLAIR hyperintense edematous regions were segmented manually. A 3D-U-Net was trained on a cohort of 260 patients from two centers to segment the GTV and the surrounding FLAIR hyperintense region. During training varying degrees of data augmentation were applied. Model validation was performed using an independent international multicenter test cohort (n = 88) including four centers. RESULTS: Our proposed U-Net reached a mean overall Dice similarity coefficient (DSC) of 0.92 ± 0.08 and a mean individual metastasis-wise DSC of 0.89 ± 0.11 in the external test cohort for GTV segmentation. Data augmentation improved the segmentation performance significantly. Detection of brain metastases was effective with a mean F1-Score of 0.93 ± 0.16. The model performance was stable independent of the center (p = 0.3). There was no correlation between metastasis volume and DSC (Pearson correlation coefficient 0.07). CONCLUSION: Reliable automated segmentation of brain metastases with neural networks is possible and may support radiotherapy planning by providing more objective GTV definitions.
Assuntos
Neoplasias Encefálicas , Radiocirurgia , Humanos , Redes Neurais de Computação , Imageamento por Ressonância Magnética , Neoplasias Encefálicas/diagnóstico por imagem , Neoplasias Encefálicas/radioterapia , Planejamento da Radioterapia Assistida por Computador , Processamento de Imagem Assistida por ComputadorRESUMO
Three-dimensional Surface Imaging (3DSI) has become a valuable tool for planning and documenting surgical procedures. Although surface scanners have allowed for a better understanding of breast shape, size, and asymmetry during patient consultation, its use has not been included in intraoperative assessment so far. Validation of the reliability of the intraoperative use of a portable handheld 3DSI equipment as a tool to evaluate morphological changes during breast augmentation surgery. The patients who underwent bilateral subpectoral breast augmentation through an inframammary incision were included in this study. Intraoperative 3DSI was performed with the Artec Eva device, allowing for visualization of the surgical area before incision, after use of breast sizers and implant, and after wound closure. Intraoperatively manual measurements of breast distances and volume changes due to known sizer and implant volumes were in comparison with digital measurements calculated from 3DSI of the surgical area. Bilateral breasts of 40 patients were 3D photographed before incision and after suture successfully. A further 108 implant sizer uses were digitally documented. There was no significant difference between manual tape measurement and digital breast distance measurement. Pre- to postoperative 3D volume change showed no significant difference to the known sizer and implant volume.
RESUMO
PURPOSE: Surgical treatment of jumper's fractures is a highly demanding situation for the surgeon due to its rareness and frequent association with severe concomitant injuries. There is no current consensus regarding a standard treatment approach, thus reducing quality of care. Our objectives were to describe, apply and assess a novel surgical technic. METHODS: The presented research is an observational retrospective study of patients who underwent the described novel surgical intervention in a level 1 trauma center. We conducted analyses of the patient cohort using patient-related outcome measures at least 1 year after surgery, as well as investigating pain, quality of life and the clinical effectiveness of the procedure. RESULTS: A total of 24 patients (17 male and 7 female) with an average age 47 ± 16.3 years were included. ISS scores ranged from 9 to 66 with a mean ISS of 40 ± 15. Clinical scores exist of 15 out of 24 patients (62.5%). The mean VAS score was 53.7 ± 12.9. The mean EQ-5D index was 0.68 ± 0.22. Significant negative correlation existed between the ISS value and the EQ-5D index (r = - 0.704; p < 0.005) and EQ-5D VAS (r = - 0.809; p < 0.001). Anatomical reduction was achieved in all patients (n = 24). Radiological follow-up was performed in 58%. CONCLUSION: We present one of the largest studies with operatively treated jumper's fractures of the sacrum. The technique is capable of reproducibly restoring the physiological anatomy of the patient and allows pain-adapted mobilization.
Assuntos
Fraturas Ósseas , Sacro , Adulto , Feminino , Fixação Interna de Fraturas/métodos , Humanos , Masculino , Pessoa de Meia-Idade , Qualidade de Vida , Estudos Retrospectivos , Sacro/diagnóstico por imagem , Sacro/lesões , Sacro/cirurgiaRESUMO
BACKGROUND: Three-dimensional Surface Imaging (3DSI) is a well-established method to objectively monitor morphological changes in the female breast in the field of plastic surgery. In contrast, in radiation oncology we are still missing effective tools, which can objectively and reproducibly assess and document adverse events in breast cancer radiotherapy within the framework of clinical studies. The aim of the present study was to apply structured-light technology as a non-invasive and objective approach for the documentation of cosmetic outcome and early effects of breast radiotherapy as a proof of principle. METHODS: Weekly 3DSI images of patients receiving either conventionally fractionated radiation treatment (CF-RT) or hypofractionated radiation treatment (HF-RT) were acquired during the radiotherapy treatment and clinical follow-up. The portable Artec Eva scanner (Artec 3D Inc., Luxembourg) recorded 3D surface images for the analysis of breast volumes and changes in skin appearance. Statistical analysis compared the impact of the two different fractionation regimens and the differences between the treated and the contralateral healthy breast. RESULTS: Overall, 38 patients and a total of 214 breast imaging sessions were analysed. Patients receiving CF-RT showed a significantly higher frequency of breast erythema compared to HF-RT (93.3% versus 34.8%, p = 0.003) during all observed imaging sessions. Moreover, we found a statistically significant (p < 0.05) volumetric increase of the treated breast of the entire cohort between baseline (379 ± 196 mL) and follow-up imaging at 3 months (437 ± 224 mL), as well as from week 3 of radiotherapy (391 ± 198 mL) to follow-up imaging. In both subgroups of patients undergoing either CF-RT or HF-RT, there was a statistically significant increase (p < 0.05) in breast volumes between baseline and 3 months follow-up. There were no statistically significant skin or volumetric changes of the untreated healthy breasts. CONCLUSIONS: This is the first study utilizing 3D structured-light technology as a non-invasive and objective approach for the documentation of patients receiving breast radiotherapy. 3DSI offers potential as a non-invasive tool to objectively and precisely monitor the female breast in a radiooncological setting, allowing clinicians to objectively distinguish outcomes of different therapy modalities.
Assuntos
Neoplasias da Mama/radioterapia , Imageamento Tridimensional/métodos , Adulto , Idoso , Idoso de 80 Anos ou mais , Neoplasias da Mama/patologia , Fracionamento da Dose de Radiação , Feminino , Humanos , Pessoa de Meia-Idade , Hipofracionamento da Dose de RadiaçãoRESUMO
BACKGROUND: The aim of this study was to compare accuracy and timing of two handheld, mobile three-dimensional surface imaging (3DSI) devices against an established non-portable medical imaging system, and to evaluate future intraoperative use for facial surgery. METHODS: Surface-to-Surface root mean square analysis was used to evaluate both a consumer device (Sense 3D) and a professional surface scanner (Artec Eva) against a reference imaging system (Vectra XT). Two assessors repeatedly 3D-imaged the facial region of an imaging phantom and 30 volunteers in two separate sessions. Using both mobile devices, intraoperative 3DSI of 10 rhinoplasty patients was compared with preoperative reference imaging. Intraclass Correlation Coefficient was calculated for repeated measurements. RESULTS: Artec Eva yielded mean deviations below 0.5 mm for the whole face and all subunits excluding the eye region. Sense 3D showed similar deviations for the whole face, but otherwise only in the central and lateral forehead unit and the medial cheek. Variability was low for both the non-portable Vectra XT and Artec Eva, whereas full-face assessment using Sense 3D resulted in high variability. When compared to the preoperative reference images, intraoperative rhinoplasty 3DSI revealed low deviations for Artec Eva and high deviations for Sense 3D. CONCLUSIONS: The 3D surfaces captured by Artec Eva showed a similarly desirable accuracy for facial imaging as Vectra XT reference images. This handheld device presents a suitable option for the objective documentation during rhinoplasty surgery. Sense 3D was unable to accurately capture complex facial surfaces and is therefore limited in its usefulness for intraoperative 3DSI.
Assuntos
Face/diagnóstico por imagem , Fotogrametria/instrumentação , Adulto , Estudos de Casos e Controles , Bochecha/diagnóstico por imagem , Desenho de Equipamento , Face/cirurgia , Feminino , Testa/diagnóstico por imagem , Voluntários Saudáveis , Humanos , Imageamento Tridimensional/instrumentação , Cuidados Intraoperatórios/métodos , Masculino , Manequins , Imagens de Fantasmas , Rinoplastia/métodosRESUMO
BACKGROUND: Three-dimensional surface imaging (3DSI) has proven to be useful in providing objective aid to the planning process and documentation of various plastic-surgical procedures. Although this technology is routinely used in the surface and volume analysis of the face and breast, it has been of limited use in registering and quantifying the resulting changes to the entire body surface. The aim of this study was the clinical implementation of 360° whole-body scans to evaluate the treatment of lipoedema. PATIENTS, MATERIALS AND METHODS: Three 3DSI devices (Eva, Thor and Sense) were tested for precision and reproducibility regarding whole-body scans. Using a standardised setup consisting of an automatic turntable and predetermined body poses, human subjects were analysed by measuring axial circumferences and quantifying defined layers of body volume. The relevant statistical and clinical deviations were subsequently evaluated. RESULTS: A standardised procedure for 360° scans was successfully implemented. All tested scanners yielded sufficient results with respect to intraindividual reproducibility (p > 0.05). The Eva and Thor scanners delivered comparable results for axial circumference and volume analysis (p > 0.05). The Sense scanner allowed for a precise analysis in the area of the body trunk, but had significant deficits regarding the lower extremity (p < 0.05). The data analysis was then successfully applied to selected clinical cases. CONCLUSION: A procedure to reproducibly capture and analyse the human body was successfully established for clinical use in plastic surgery. Two of the tested 3DSI devices allowed for an objective surface and volume analysis of the human body. The third scanner (Sense) offered the ability to perform 360° scans at a low cost, albeit lacking in precision when applied to certain areas of the body. These findings may help to objectively evaluate the effects of different procedures on the entirety of the body surface in the future.
Assuntos
Corpo Humano , Imageamento Tridimensional , Imagem Corporal Total , Extremidades , Humanos , Reprodutibilidade dos TestesRESUMO
BACKGROUND: Soft-tissue filler injections for the treatment of facial aging can result in different skin surface effects depending on the targeted facial fat compartment and fascial plane. This work investigates the tissue response of defined amounts of soft-tissue filler material injected into superficial and deep facial fat compartments by means of the calculation of the surface-volume coefficient. METHODS: Four fresh frozen cephalic specimens obtained from human donors (three female and one male; mean age, 74.96 ± 22.6 years; mean body mass index, 21.82 ± 6.3 kg/m) were studied. The superficial and deep lateral forehead compartments, deep temporal fat pad, sub-orbicularis oculi fat compartment, and deep medial cheek fat compartment were injected with aliquots of 0.1 cc of contrast-enhanced material and scanned using three-dimensional surface imaging, resulting in a total of 226 injection and scanning procedures. RESULTS: The sub-orbicularis oculi fat compartment revealed the highest correlation coefficient (rp = 0.992; p < 0.001) and the highest surface-volume coefficient (0.94). The compartment with the lowest tissue response was the deep medial cheek fat compartment (rp = 0.745; p < 0.001; surface-volume coefficient = 0.29), followed by the deep lateral forehead compartment (rp = 0.814; p < 0.001; surface-volume coefficient = 0.68), superficial lateral forehead compartment (rp = 0.824; p < 0.001; surface-volume coefficient = 0.74), and deep temporal fat pad (rp = 0.947; p < 0.001; overall surface-volume coefficient = 0.64). CONCLUSIONS: These results, confirmatory in their nature to current injection strategies, provide evidence for the validity and reliability of the surface-volume coefficient. Injection procedures should be targeted in terms of facial fat compartments and fascial planes for a desired aesthetic outcome, as each fat compartment and fascial plane has unique tissue responses to injected soft-tissue fillers.
Assuntos
Tomografia Computadorizada de Feixe Cônico/métodos , Imageamento Tridimensional , Tecido Adiposo/anatomia & histologia , Tecido Adiposo/diagnóstico por imagem , Idoso , Idoso de 80 Anos ou mais , Cadáver , Preenchedores Dérmicos , Dissecação , Face/anatomia & histologia , Face/diagnóstico por imagem , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Sensibilidade e Especificidade , Envelhecimento da Pele , Gordura Subcutânea/anatomia & histologia , Gordura Subcutânea/diagnóstico por imagemRESUMO
BACKGROUND: Three-dimensional surface imaging (3DSI) has shown promise for plastic surgeons to objectively assess changes in body contour and breast volume. OBJECTIVES: To assess the surgical outcome after bilateral subcutaneous mastectomy (BSM) and water jet-assisted liposuction (WAL) as treatment for idiopathic gynecomastia, using 3DSI to document changes regarding nipple-areolar complex (NAC) and breast volume. METHODS: Thirty male patients (Simon II A to B) receiving BSM and WAL were enrolled. Eight subjects received additional mastopexy and NAC reduction. Use of a Vectra 3D Imaging System® before and 6 months after surgery provided data regarding changes of NAC placement and NAC and breast dimensions. The sum volume of intraoperatively approximated lipoaspirate before and after centrifugation and mastectomy specimens quantified using water displacement were compared with 3D-assessed differences in volume. RESULTS: When compared to the NAC dimensions defined during surgery, patients receiving NAC reductions showed nonsignificant postoperative changes in NAC dimensions. Patients without additional mastopexy showed a significant (p < 0.001) vertical (15.7 ± 14.3%) and horizontal (17.1 ± 15.0%) reduction in NAC diameter. 3D volume changes (92.8 ± 26.4 mL) showed significant differences (p < 0.001) 6 months after surgery compared to the intraoperatively measured lipoaspirate before (182 ± 54.5 mL) and after (120 ± 34.6 mL) centrifugation. CONCLUSION: Although it is clear that patient satisfaction must remain the prime focus of surgical body contouring, 3DSI has proven valuable to objectively demonstrate both the anticipated outcome and further findings regarding treatment of idiopathic gynecomastia. LEVEL OF EVIDENCE IV: This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266 .
Assuntos
Ginecomastia/diagnóstico por imagem , Ginecomastia/cirurgia , Imageamento Tridimensional , Lipectomia , Mastectomia , Adolescente , Adulto , Ginecomastia/patologia , Humanos , Lipectomia/métodos , Masculino , Mastectomia/métodos , Pessoa de Meia-Idade , Mamilos/anatomia & histologia , Mamilos/diagnóstico por imagem , Tamanho do Órgão , Resultado do Tratamento , Adulto JovemRESUMO
BACKGROUND: Handheld 3-dimensional surface imaging (3DSI) devices of various precision are becoming more versatile in their applications and more widely accepted by clinicians for documentation. OBJECTIVES: The authors sought to compare the precision of facial volumetric change measurements of 3 3DSI devices in the cadaveric model: Eva (Artec 3D Inc., Luxembourg), Sense (3D Systems, Rock Hill, SC), and iSense (3D Systems, Rock Hill, SC). METHODS: A total of 336 scanning and analysis procedures were carried out in 4 cephalic specimens (mean age, 77.25 ± 24.3 years; mean BMI, 21.76 ± 6.6 kg/m2). Two superficial and 2 supraperiosteal regions of interest were injected with 0.5-cc aliquots and subsequently scanned using the 3 different scanners. Correlation coefficients between the injected and measured volume were computed. RESULTS: The correlation coefficient for the Eva scanner was for subcutaneous regions of interest rp = 0.935 and for the supraperiosteal regions of interest rp = 0.966, compared with rp = 0.760 and rp = 0.364 (superficial vs supraperiosteal) for the Sense and rp = 0.694 and rp = 0.382 (superficial vs supraperiosteal) for the iSense scanner. CONCLUSIONS: 3DSI devices are capable of measuring surface volume changes of the face at a level of 0.5-cc surface volume change and can thus be regarded as useful tools in the preinterventional, intrainterventional, and postinterventional phases of a treatment. One of the 3 evaluated scanners provided very high correlation coefficients between the injected and the measured volume (Eva), whereas the other evaluated 3DSI devices provided moderate (Sense) and low (iSense) coefficients.