3D imaging and printing for structural heart diseases

Ph.D.Structural heart diseases (SHD) refer to a variety of non-coronary structural diseases of the cardiovascular system including valvular heart diseases, congenital heart diseases, and pathologies of the great vessels. In the past 5 years, there has been a significant increase in the types of percutaneous catheter-based devices and interventions developed for various SHD. A wide clinical adoption of these percutaneous procedures, however, depends heavily on outcome of these procedures relative to conventional treatment such as open surgery. The procedural success of SHD interventions requires careful planning and understanding of the anatomy and the interactions between the anatomical structures of interest and the implanted devices. Precise knowledge of the structures of interest is crucial for correct sizing and safe placement of interventional devices. However, complexity and variability of cardiovascular structures exist and these anatomical variations among individuals pose challenges to achieving the optimal procedural outcome in all individuals. Conventional cardiac imaging techniques such as two-dimensional (2D) echocardiography and fluoroscopy have limited capability in unveiling the spatial relations of anatomical structures. Even with advanced imaging techniques such as three-dimensional (3D) echocardiography and computed-tomography (CT), complete understanding of 3D geometry remains challenging, especially when it comes to complicated cases. There is a pressing need in developing a clinical tool for optimizing the procedural planning and reducing the learning curve of the interventionists.To this end, the introduction of 3D printing technology becomes a potential game-changer. 3D printing refers to the manufacturing process by which physical objects are built by depositing materials in layers based on digital models. Medical 3D printing is the fabrication of anatomical structures, typically derived from volumetric medical image data sets. 3D printing can be applied in cardiovascular medicine by creating patient-specific models of any given cardiovascular structure, using source data derived from CT, magnetic resonance imaging (MRI), or 3D echocardiography. High-resolution, high-quality reconstructed images (less than 1 mm thin) obtained by advanced imaging allow detailed, accurate anatomy delineation that is represented on the 3D-printed models. These 3D-printed models enable visual inspection and physically manipulated, providing real volumetric assessment and unparalleled tactile perception of complex cardiovascular pathology. Moreover, simulated interventional procedure and device testing can be conducted on the 3D model for personalized procedural planning.Despite the promising potentials of 3D printing in SHD intervention, widespread adoption of 3D printing has been limited by the lack of robust evidence that systemically demonstrates effectiveness, the lack of validated technical standard, and potentially increased costs. The accuracy and reproducibility of cardiac structural geometry must be validated across different source imaging modalities, for various 3D printing materials and methods, and clinical benefits in improving patient outcomes need to be demonstrated in various interventional scenarios.This thesis aimed to 1) define and validate the suitable 3D printing methodology including the use of material with physical properties that can closely mimic cardiovascular tissues (Study 1); 2) explore the feasibility and clinical value of 3D printing in various SHD interventions, including the impact on advanced pre-operational planning, decision making on device choice and appropriate sizing, and accuracy in prediction of clinical outcomes (Study 2-5).Study 1 compares the mechanical properties of 3D printing materials and human tissues in vitro. Preconditioned fresh samples from patients and 3D printed samples were loaded accordingly for vertical uniaxial tensile testing to validate the material properties against human tissues.The clinical studies were therefore conducted with 3D printing material validated in Study 1. Clinical application of patient specific 3D-printed models in following groups of patients was studied:1) Application in LAA occlusion I- retrospective derivative cohort (72 patients, Study 2).2) Application in LAA occlusion II- prospective validation cohort (32 patients, Study 3).3) Application in investigation of mechanism for coronary occlusion post transcatheter aortic valvular implantation (TAVI) - case study (1 patient, Study 4).4) Application in atrial septal defect (ASD) closure - case study (1 patient, Study 5).Study 2 evaluated the feasibility and accuracy of echo-based 3D-printed models in the prediction of size and position of interventional device. Importantly, this study assessed the association of model-based device selection with procedural safety and efficacy and determined if pre-procedural model testing leads to superior outcome. In 72 patients who underwent LAA occlusion under 3D-TEE, 3D models of the LAA were created form 3D datasets retrospectively (retrospective cohort). The optimal device determined by in vitro model testing was compared to the actual device used. Association of model-match/mismatch device with outcomes was analysed. Study 3 is an extension of Study 2 with 32 patients prospectively involved. Device selection was guided by 3D models in adjunct to 3D-TEE (prospective cohort). Impact of model-sizing on outcomes was assessed by comparing the 2 cohorts. The main findings of these 2 studies were: 1) Patients with model-mismatch sizing had more procedure-related adverse events over 3.0±2.3 years of follow-up. 2) Compared to the retrospective imaging-guided cohort, the prospective model-guided patients achieve higher implant success with no procedural complications, shorter procedure time and less device used per procedure. 3) High agreement of compression and protrusion was unveiled between procedural and testing results. 4) Predictors for sizing mismatch were non-windsock morphology and prominent trabeculations near device landing zone.Our initial successful experience of using 3D printing in LAA occlusion opens up opportunities for investigation in other SHD interventions. Study 4 reported the performance of 3D printing in investigation of the mechanism for coronary occlusion post-TAVI. The patient was assessed as low-risk for coronary occlusion because pre-TAVI CT images showed adequate coronary ostium height and sinuses of Valsalva width. Unfortunately, the following procedure was complicated with left coronary artery occlusion 12 hours post-TAVI. To investigate the mechanism of coronary occlusion, a patient-specific aortic root model based on pre-TAVI CT was therefore generated. Regarding the different cardiac components, multi-material printing was used to precisely portray the flexible aortic root superstructure and the rigid calcified leaflets on the model. A 26mm transcatheter heart valve (THV) (CoreValve; Medtronic, Ireland) was deployed onto the model, followed by post-implant CT. It showed from the THV/aortic root model that the left coronary ostium was occluded by the calcified cusps crushed against the wall of the sinuses of Valsalva after THV implantation.Study 5 presents a case of 3D printing as a clinically applicable heart modelling technology for pre-procedural planning of a complex SHD intervention. The patient had a history of ASD closure but found with peri-device leak. The interventional cardiologists were concerned that conventional procedural method with commonly used round occluder might have chance to fail to cover the crescent-shaped leak. Inappropriate device would interfere the important surroundings such as superior vena cava. A 3D-printed model segmented from 3D-TEE was therefore created using multi-material 3D printing techniques. Several devices were tested on the model, including a vascular plug, a patent foramen ovale (PFO) occlude, and ASD occluders of different sizes. Most of them failed to achieve ideal procedural results except one type and size of ASD occluder. The actual procedure was successfully performed with suggested device and with excellent agreement to the simulation.In conclusion, research with phantoms produced by 3D printing can help to elucidate physiological processes that are not yet fully understood along with a better understanding of complex pathologies. Patient-specific 3D modelling offers incremental outcome-relevant information for SHD interventions. Application of 3D printing may directly impact our ability to select appropriate patients and devices for SHD interventions, anticipate procedural complications, and potentially improve safety and efficacy of the procedure especially in complex anatomy.結構性心臟病泛指一大類先天性或獲得性的以心臟和大血管結構異常為主要表現的心血管疾病，如傳統定義的先天性心臟病和心臟瓣膜病等。介入治療已成為結構性心臟病患者治療的主要手段。在過去五年中，行介入治療的機構性心髒病患者人數增長迅速。然而，介入手術的推廣很大程度上在於其臨床療效是否優於傳統手術。詳細的術前準備和評估是手術成功的關鍵。術前準備主要包括對患者解剖結構的了解和預測介入術後植入物和周圍組織的關係。精準的解剖知識和手術經驗是不可或缺的必備因素。但是，患者的情況千變萬化，不可同一而論，要想取得理想的手術結果是一件十分有難度的事情。介入手術十分依賴於影像學，然而傳統的影像學技術，例如二維超聲和X線透視無法提供精確的、立體的解剖信息。雖然影像技術的發展稍可彌補傳統技術的缺點，但遭遇複雜案例時，目前的影像手段仍稍顯不足。因此，我們迫切需要借助其他臨床工具來優化術前準備，及縮短介入醫生的學習週期。三維打印技術的發展可能可以改變這個局面。三維打印技術又称增材制造、積層製造，是一種通過逐層堆積的方式來構造物體的技術。三維技術在醫學領域主要利用容積三維影像資料，快速重建和製作人體器官模型。在心血管專業，三維技術利用計算機斷層掃描、心臟磁共振或者超聲的數據，重建心臟模型。利用先進影像技術所採集到的高分辨率、高質量的圖像可以提供精度小於1毫米的解剖細節，並呈現在三維模型上。三維模型可以直接反應複雜的心血管空間形態，術者也可以直接在模型上進行測量和手術模擬操作。但是，三維技術的推廣受限於缺乏系統的循證醫學證據，統一有效的技術標準和較高的成本。不同的源數據、打印材料和臨床應用都需要經過驗證才能證明三維打印心臟打印模型的準確性。因此，此論文的目的在於：（1）比較和驗證三維打印材料和心臟組織之間的機械性能（研究1）；（2）探討三維打印在結構性心髒病方面的可行性和應用價值，主要包括三方面：術前評估、封堵器型號和大小的選擇，預測臨床療效（研究2-5）。研究1是一個體外驗證實驗。我們通過垂直單軸拉伸實驗，比較6例新鮮的病人左心耳和不同的三維材料之間的機械性能。結果顯示，在生理應變範圍內，材料Agilus 30的物理特性與人體左心耳最為接近。在驗證試驗的基礎上，我們開展了一系列關於三維打印的臨床應用研究，包括：（1） 回顧性研究三維打印在左心耳封堵術中的應用價值（研究2，共納入72例患者）（2） 前瞻性研究三維打印在左心耳封堵術中的應用價值（研究3，共納入32例患者）（3） 運用三維打印研究經導管主動脈瓣植入術後並發冠狀動脈閉塞的機制（案例研究）（4） 運用三維打印模型模擬房間隔封堵術後殘餘漏的封堵（案例研究）研究2是一個回顧性研究，評價基於食道超聲的三維打印模型在左心耳封堵術中，對於左心耳封堵器大小及位置的預測價值。此外，這是首個將模型的預測結果和介入手術的有效性和安全性作相關分析的研究，觀察三維模型的預測結果是否與較好的臨床轉歸相關。本研究共回顧性納入72例行左心耳封堵術的患者，將患者的術前三維食道超聲數據作為源數據，製作患者的左心耳模型。我們根據患者實際使用的封堵器型號和三維模型的測試型號是否一致將患者分成兩組，一組患者實際使用的封堵器型號與三維打印模型的測試結果一致，另一組患者則結果不一致，並分析比較兩組的臨床轉歸。研究3是在研究2基礎上的擴展研究。研究3前瞻性地納入了32例行左心耳封堵術的患者。三維打印模型作為術前參考的標準之一，指導術中封堵器的選擇。研究2（前瞻性隊列研究）主要採用食道超聲作為參考，選擇封堵器型號；而研究3（回顧性隊列研究）加入了三維打印模型的預測結果作為術前參考，選擇封堵器型號。因此，通過比較兩個隊列的臨床效果，可以得出哪種方法指導介入手術更有優勢。研究發現：（1）在3.0±2.3年的隨訪中，實際所用封堵器與模型預測不一致的患者有更多的臨床不良事件；（2）和回顧性隊列相比，前瞻性隊列的患者成功率更高，無相關併發症，手術時間更短，所用封堵器的個數更少；（3）手術和模型預測的封堵器露肩值和壓縮比一致性高；（4）導致兩種方法結果不一致的影響因子主要有非風向標型左心耳，和左心耳的近端錨定區有粗大的肌小梁。有了三維打印在左心耳封堵術中的經驗，我們將此應用拓展到其他結構性心髒病中。研究4是一個案例分析，報道了一起經皮主動脈瓣置換術後並發冠狀動脈閉塞的案例。我們利用三維模型回顧性模擬手術過程，尋找發生併發症的可能原因。術前的影像學資料顯示該患者無論冠狀動脈開口高度和主動脈瓣竇寬度都足夠，因此發生冠狀動脈閉塞的可能性較低。然而，術後12小時之內就發生了冠狀動脈閉塞。因此，爲了探討冠狀動脈閉塞的發生機制，我們運用了復合材料打印的方法，製作了這個個體化的主動脈根部模型，目的為了通過不同材料打印不同的心臟組織，如柔軟的主動脈結構和堅硬的主動脈鈣化，能更真實地展現心臟結構。我們仿照原本的手術過程，將直徑26毫米的人工主動脈瓣放置在主動脈模型裡，然後用CT掃描模型並加以分析，發現出現經皮主動脈瓣植入術後發生冠狀動脈閉塞的可能原因是鈣化增厚的主動脈瓣受到人工植入瓣的擠壓，阻塞了冠狀動脈入口。研究5也是一個案例研究，報道了在三維打印模型上進行一例複雜結構性心臟病介入術的術前模擬。本案例中的患者既往有房間隔缺損封堵術的病史，但在術後6個月發現了半月牙型的殘餘漏。我們擔心常用的圓型封堵器有可能無法完全封堵殘餘漏，亦有可能影響周邊的結構，如阻塞下腔靜脈。因此，我們採用復合材料打印的方法，製作了一個基與食道超聲的房缺+封堵器的殘餘漏模型。我們在模型上試驗了不同型號和類型的封堵器，包括卵圓孔未閉封堵器、血管封堵器和房間隔缺損封堵器，最後發現一個合適大小的房間隔缺損封堵器可以達到理想的手術要求。手術結果和模擬結果完全一致。結論：三維打印的人體器官模型有助於闡明尚未完全理解的生理過程，以及加強對複雜病理的理解。個體化的三維模型為結構性心臟病介入手術提供了額外的空間信息。在介入手術中應用三維打印可能會幫助我們篩選合適的患者、選擇適當的封堵器、預測手術併發症，特別在複雜的手術中可能幫助術者提高手術的安全性和有效性。Fan, Yiting.Thesis Ph.D. Chinese University of Hong Kong 2019.Includes bibliographical references (leaves 106-112).Abstracts also in Chinese.Title from PDF title page (viewed on 20, November, 2020)