Summary Background and rationale The anterior cruciate ligament (ACL) is a ligament of the knee joint that has a crucial role in determining its antero-posterior and rotational mechanical stability. The ACL originates at the posterior part of the lateral wall of the femoral intercondylar notch and descends until its tibial insertion, nearly in the middle of the tibial plateau. From a functional point of view, two families of fibres can be identified in it: the anteromedial (AM) and the posterolateral (PL) bundles. Each of them has a different elongation pattern during knee flexion: the AMB stretches in flexion, while the PLB in extension. The ACL is the most injured ligament of the human body and injuries occur mainly in young and sports-active subjects. The incidence of ACL tears is estimated to be between 1 and 10 in 1000 people per year, with the total number being between 100.000 and 200.000 in the United States. Due to its morphology, ACL tears are mainly caused by a high valgus stress to the extended knee and secondary by hyperextension or extreme internal rotation of the tibia. Given the fundamental role of the ACL in the biomechanics of the knee, its rupture leads to general mechanical instability and abnormal kinematics of the joint, which speeds up the osteoarthritis (OA) progression. The goal standard treatment for ACL tear is the surgical arthroscopic reconstruction (ACLR), whose goal is to restore the pre-injury kinematics of the knee. The ligament is substituted with a graft, typically harvested from the patient's hamstring or patellar tendon, which is fixated in tunnels drilled into the femur and the tibia. The efficacy of the surgical act in restoring the kinematics of the knee has been proven to be highly sensitive to the location where the tunnels are placed and the native footprint (namely, the anatomical point) has been described as the optimal position. However, the original insertion sites are difficult to be determined during the arthroscopical intervention. Indeed, 25-88% of ACL reconstruction failures have been attributed to femoral tunnel misplacement. Identification of the ACL femoral footprint from endoscopic visualization is often not possible and the estimation of its location on the basis of average anatomical guidelines does not give accurate results because of the morphological differences between individuals. Furthermore, accurate execution of the pre-operative plan has been shown to be difficult to achieve even for experienced surgeons, with reported errors between 3 and 12 mm. Computer aided surgery has allowed to reach higher levels of accuracy in tunnel placement with respect to free-hand drilling. However, this technique has seen limited diffusion, since it is still costly in terms of equipment and time. Patient-specific instrumentation has been designed and successfully used in other orthopaedics branches by coupling 3D modelling from clinical images and 3D printing rapid prototyping. Personalized guides seemed to be an inexpensive method to address an issue that is so strongly related to the subject’s anatomical characteristics. We hypothesized that accurate femoral tunnel placement could be achieved with a patient-specific surgical guide. The goal of this thesis was to perform a proof of concept of customized 3D printed molds as a method for accurate femoral tunnel placement that would adapt to the current clinical and surgical practice without further significant costs. A guide that could uniquely fit the patient's anatomy would be an ideal intraoperative tool to translate the preoperative plan (i.e. any concept for femoral tunnel placement) into the actual surgery. This will allow patients to receive truly personalized (anatomical, ideally) ACL reconstructions, regardless the surgeon's expertise, and decrease the incidence of graft failure, thanks to a correct tunnel placement. If coupled with an appropriate method for preoperative identification (or precise estimate) of the native femoral footprint, this will finally allow accurate anatomical tunnel placement, even for all those surgeons that perform only a small number of ACLR every year, since 70% of the ACLR are performed by orthopaedic surgeons that have an experience as low as 20 reconstructive procedures per year). Our research question was: are 3D printed patient-specific surgical guides a feasible solution to reach accurate femoral tunnel placement in anterior cruciate ligament reconstruction? Materials and methods Different designs of 3D-printed patient-specific guides for transferring a templated femoral tunnel position into surgery were developed and evaluated following the standard engineering design process. Overall, three phases of prototype development and evaluation were carried out, each of them addressing the verification of specific design requirements. Design requirements The goal was to develop a device that would uniquely fit the surface of the patient's femur and whose main characteristic would be to accurately report into the surgery the position of the femoral tunnel centre as pre-operatively planned. Different design requirements were set for the device to answer the research question. First, the device should be compatible with arthroscopic surgery and be therefore insertable through a 20 x 10 mm incision. Secondly, the guide should be superior to free-hand tunnel positioning in terms of accuracy: 5 mm of maximum error was set as a reference value. In view of a possible commercialization, the developed guide should be sterilisable and comply with the regulatory standards to certify its biological safety for both patients and users. To adapt as much as possible to the current clinical practice, the design of the guide should, ideally, make use of the standard clinical imaging for ACLR, being 2D MRI. Two characteristics will be then needed by the guide to comply with its final use: fitting, namely perfect congruence to the surface of the femur, and stability, to avoid any slipping on the bone during manipulation and drilling. This could be achieved by selecting an appropriate combination of bony and cartilaginous features to constrain the motion of the guide. Development workflow Three generations of guides were developed with similar procedures. The first preliminary concept was designed on a commercial sawbone and differed from the definitive workflow in terms of imaging technique. The second and third generations more closely reproduced the clinical scenario and were designed for a (single) human cadaver knee. The intended final workflow for the design of the guides consisted in four steps: clinical image acquisition of the patient’s injured knee, segmentation of the images to obtain a 3D model of the distal femur, identification of the intended location for the femoral tunnel, design of the mold with the internal surface coincident to that of the femur for perfect fit. CT (computed tomography) scans were used for the sawbone, since higher contrast is needed for a synthetic femur. 3D isotropic MRI scans were used for the cadaver knee, as the resolution of the 2D MRI imaging clinical standard was found to be insufficient. To evaluate the accuracy of the guides in translating into the surgical act a pre-operatively planned femoral tunnel position, a 2 mm cylindrical guide-hole was reported in the lateral wall of the intercondylar notch of each guide in the defined position. Although determining an estimate of the actual location of the ACL footprint was not among the goals of this project, a realistic value was calculated by applying guidelines from the literature. The apex of the deep cartilage (ADC) was chosen as anatomical landmark to give an estimate of the centre of the AMB (used as a reference point for the second generation of guides) and of the ACL footprint (used as a reference point for the third generation of guides). Three different design phases were carried out, each of them resulted in one generation of prototypes followed by specific experimental tests to evaluate adherence with design requirements. The outputs of each test session were used as inputs for the following design phase. The moulds were created on the 3D surface model of the femur to perfectly fit the surface of the condyles as acquired and segmented from the clinical images. Different bony and cartilaginous anatomical features were explored to guarantee adequate stability of the devices once positioned. Analysis methods After manufacturing, each generation of guides was tested and graded according to its development stage and design requirements. The first generation was evaluated on the sawbone, with regards to two characteristics: stability and fitting. The first requirement was verified through manual inspection. Then, a structured light 3D scanner was used to quantify possible mismatches between the surfaces of sawbone and guides that would have led to fitting issues, and to identify their origin along the design process. The second and third generations were instead tested in the final scenario: the guides were employed during an ACLR surgery performed in a cadaveric knee. The qualitative performances of the third generation of guides was judged by a design attribute assessment. The main functional characteristics of the guides were expressed by three attributes (positioning, stability and ease of insertion) and the two surgeons were asked to assign to each of them a score from 1 to 5 (from lowest to highest). In the second cadaver test the accuracy of the last two generations of guides was evaluated on the dissected femur with a 3D optical tracker. Results Divided into three generations, a total of 10 prototypes were developed. Polyamide 12 was chosen as manufacturing material because of its certified biocompatibility and good mechanical properties. First generation: sawbone evaluation and surface scan Three prototypes were designed to achieve fitting to the sawbone using the medial cartilage edge of the lateral condyle as primary kinematic constraint. Cartilage coverage, namely the extent of the guides on the cartilage of the lateral condyle, was the main variable parameter being investigated. The middle guide coupled intermediate size (two of its dimensions did not exceed 2 cm), satisfactory stability and good ease of positioning, closely meeting the design requirements and making it the best prototype for further development. The manufacturing error was generally comparable to the tolerance of 0.1 mm provided by the manufacturer. Second generation and first cadaver test As the sawbone test showed that good stability was guaranteed by the employed concept and the fitting reached was appropriate, four new molds were created based on the 3D surface model of the cadaveric femur following the same concept. In addition to the medial cartilage edge of the lateral condyle, two other anatomical structures in the distal femur were explored for additional kinematic constraint in the other three prototypes: the most anterior part of the medial cartilage edge of the lateral condyle (the roof of the notch) and the apex of the deep cartilage (ADC). The handling and insertion mechanism was also further developed to suit the real surgical scenario of the cadaver evaluation. A standalone handle was manufactured with the dimensions of a typical arthroscopical instrument and a squared-taper mechanism was designed for the coupling with the guides. All of the guides could be straight-forward positioned in the intended position without hesitation. Once placed, each guide proved an adequate degree of stability, although the two guides that relied on the cartilage edge of the roof of the notch showed a higher constraint to rotations around the cartilage edge. The insertion and handling mechanism proved too flexible and was therefore evaluated as ineffective in facilitating a precise positioning. Third generation and second cadaver test Although the previous models provided adequate stability, the incompatibility with the arthroscopic procedure required a substantial design modification towards a thinner device. Three new concepts were developed to more strictly limit two of the three dimensions of the guides to that of the portals. The new guides were provided with a stiffer integrated cylindrical handle for both insertion and positioning from the antero-lateral portal. Every guide could be easily inserted through the antero-lateral portal, especially the thinner, “ADC 2” model. Positioning in the intended position was reached for all the guides with visual endoscopic support and facilitated by visualization of the computer-drawn models. Overall, the ADC 2 guide received the highest score, thanks to its ease of insertion and positioning. All the guides reached an accuracy higher than 5 mm. Within the second generation, a substantial difference in accuracy was observed between the models that used the most anterior part of the cartilage edge and the ones that did not. The guide with highest level of constraint reached the highest accuracy and the one with lowest constraint achieved the lowest accuracy. The same was valid also for the two guides of the third generation that had similar design. On average, the third generation outperformed the previous one. Discussion The last generation of fully functional prototypes was designed from the 3D MRI images of a cadaver knee and tested under standard surgical conditions. The guides were manufactured with in a biocompatible and sterilisable polymer, showed excellent compatibility with the arthroscopic setup and good fitting on the anatomy of the femur. The guides achieved higher accuracy than an average experienced surgeon, indicated in the literature as 3-12 mm. Thanks to the use of 3D MRI, this method for accurate femoral tunnel placement obviates radiation exposure and eliminates the need for expensive, complex and time-consuming techniques. The followed development workflow reached satisfactory fitting from the segmentation of clinical images, although it is difficult to say whether it influenced stability. Imaging standard practice was not significantly altered. Before reaching the clinical application, future research should be focused on further demonstrating the robustness of the functional concept by overcoming the methodological limitations of this study, namely sample size and inter-user reliability of measurements. Conclusions and clinical relevance The results of this study show that 3D printed femoral guides have the potential to improve the clinical outcomes of ACLR. If coupled with an effective way to identify the native anatomical footprint, the developed devices could serve as a useful tool for surgeons (especially the most unexperienced ones) to improve knee laxity and long-term clinical outcome of ACLR, like reduction of osteoarthritis incidence.
Sommario Il legamento crociato anteriore (LCA) è un legamento dell’articolazione del ginocchio che ha un compito cruciale nel determinarne la stabilità meccanica rotazionale e in direzione antero-posteriore. Il LCA origina nella regione posteriore della parete laterale del solco intercondiloideo e discende sino alla sua inserzione tibiale, in prossimità del centro del piatto tibiale. Da un punto di vista funzionale, due gruppi di fibre possono essere in esso identificate: il fascio antero-mediale e il fascio postero-laterale. Ciascuno di essi ha un diverso comportamento in allungamento durante la flessione del ginocchio: il primo si allunga in flessione, mentre il secondo in estensione. Il LCA è il legamento del corpo umano più soggetto a lesioni, che occorrono principalmente in soggetti giovani e sportivi. L’incidenza di rottura del LCA si stima essere tra 1 e 10 casi ogni 1000 persone all’anno, con circa 100.000-200.000 casi annuali nei soli Stati Uniti. Per via della sua morfologia, le rotture del LCA sono principalmente causate da elevate sollecitazioni in direzione mediale in estensione del ginocchio e, in secondo luogo, da iperestensione o eccessiva intra-rotazione della tibia. Dato il ruolo fondamentale del LCA all’interno della biomeccanica del ginocchio, la sua rottura porta all’articolazione instabilità meccanica e una cinematica anormale, che accelerano l’insorgenza di osteoartrite (OA). Il trattamento standard per la rottura del LCA è la sua ricostruzione artroscopica (RLCA), il cui obiettivo è quello di ripristinare la cinematica del ginocchio precedente alla lesione. Il legamento è sostituito da un innesto, tipicamente prelevato dal tendine del bicipite femorale o dal tendine rotuleo, che viene fissato in canali ossei creati nel femore e nella tibia. L’efficacia dell’atto chirurgico nel ripristinare la cinematica del ginocchio si è dimostrata essere altamente sensibile al luogo in cui vengono perforati i tunnel, e l’impronta nativa del LCA è stata descritta come posizione ottimale. Tuttavia, la determinazione dei luoghi di inserzione originari durante la procedura artroscopica è difficile realizzazione. Infatti, tra il 25 e l’88% dei casi di fallimento della ricostruzione del LCA sono attribuiti a una errata collocazione del tunnel femorale. L’identificazione dell’impronta femorale del LCA mediante l’endoscopio non è sempre possibile e la stima della sua posizione tramite l’impiego di linee guida basate su studi anatomici porta a risultati non accurati per via delle differenze morfologiche tra individui. Inoltre, è stato dimostrato che eseguire un piano preoperatorio in modo accurato è difficile da realizzare anche per chirurghi esperti, con errori riportati tra 3 e 12 mm. La chirurgia computer-assistita ha permesso di raggiungere livelli superiori di accuratezza nel posizionamento dei tunnel ossei rispetto alla perforazione a mano libera. Tuttavia, questa tecnica ha visto una diffusione limitata, poiché risulta ancora costosa in termini di tempo e attrezzature. La strumentazione paziente-specifica è stata progettata e utilizzata con successo in altri settori dell’ortopedia mediante l'accoppiamento della modellazione 3D con immagini cliniche e la stampa 3D. Le guide personalizzate sembrano essere un metodo economico per affrontare un problema così fortemente correlato alle caratteristiche anatomiche del soggetto. Abbiamo ipotizzato che un posizionamento accurato del tunnel femorale potesse essere ottenuto con una guida chirurgica specifica per il paziente. L'obiettivo di questa tesi è stato quello di eseguire una dimostrazione del concetto di stampi personalizzati fabbricati mediante stampa 3D come metodo per un accurato posizionamento del tunnel femorale che si adattasse all'attuale pratica clinica e chirurgica senza ulteriori costi significativi. Una guida che si adatti in modo univoco all'anatomia del paziente sarebbe uno strumento intraoperatorio ideale per tradurre il piano preoperatorio (i.e. qualsiasi concetto di posizionamento del tunnel femorale) nell'effettivo intervento chirurgico. Ciò consentirà ai pazienti di ricevere ricostruzioni del LCA realmente personalizzate (anatomicamente, nel caso ideale), indipendentemente dall'esperienza del chirurgo, e di diminuire l'incidenza del fallimento dell'innesto, grazie ad un corretto posizionamento del tunnel. Se accoppiato con un metodo appropriato per l'identificazione preoperatoria (o una stima precisa) dell'impronta femorale originaria, questo permetterà finalmente un accurato posizionamento anatomico del tunnel, anche per tutti quei chirurghi che eseguono solo un minimo numero di LCAR ogni anno, dal momento che il 70% degli interventi viene eseguito da chirurghi ortopedici che eseguono meno di 20 procedure ricostruttive all'anno). La nostra domanda di ricerca è stata: possono guide chirurgiche paziente-specifiche prodotte mediante stampa 3D rappresentare una soluzione fattibile per raggiungere un posizionamento accurato del tunnel femorale nella ricostruzione del legamento crociato anteriore? Materiali e metodi Diversi design di guide paziente-specifiche stampate in 3D per il trasferimento nell’atto chirurgico di una posizione del tunnel femorale pianificata sono stati sviluppati e valutati seguendo il processo standard di progettazione ingegneristica. Complessivamente sono state eseguite tre fasi di sviluppo e valutazione del prototipo, ciascuna delle quali ha riguardato la verifica di specifici requisiti di progettazione. Requisiti di progettazione L'obiettivo è stato quello di sviluppare un dispositivo che si adattasse in modo univoco alla superficie del femore del paziente e la cui caratteristica principale fosse quella di riportare con accuratezza nella chirurgia la posizione del centro del tunnel femorale pianificata pre-operativamente. Diversi requisiti di progettazione sono stati stabiliti per il dispositivo per rispondere alla domanda di ricerca. In primo luogo, il dispositivo deve essere compatibile con la chirurgia artroscopica ed essere quindi inseribile attraverso un'incisione di 20 x 10 mm. In secondo luogo, la guida dovrebbe essere superiore in termini di accuratezza al posizionamento del tunnel a mano libera: 5 mm di errore massimo è stato impostato come valore di riferimento. In vista di una possibile commercializzazione, la guida sviluppata deve essere sterilizzabile e conforme agli standard normativi per certificarne la sua sicurezza biologica sia per i pazienti sia per gli utenti. Per adattarsi il più possibile all'attuale pratica clinica, la progettazione della guida dovrebbe, idealmente, utilizzare gli standard clinici di imaging per la chirurgia del LCA, ovvero la risonanza magnetica nucleare (RMN) 2D. Infine, due caratteristiche saranno necessarie affinché la guida risulti adeguata al contesto finale: perfetta congruenza con la superficie del femore, e stabilità, per evitare qualsiasi scivolamento sull'osso durante la manipolazione e la perforazione. Ciò potrebbe essere ottenuto scegliendo una combinazione appropriata di caratteristiche ossee e cartilaginee per limitare il movimento della guida. Iter di progettazione Tre generazioni di guide sono state sviluppate con simili procedure. Il primo concept preliminare è stato progettato su un osso sintetico commerciale e si è differenziato dall’iter di sviluppo definitivo in termini di tecnica di imaging. La seconda e la terza generazione, invece, riproducono più da vicino lo scenario clinico e sono state progettate per un (singolo) ginocchio da donatore. Il flusso di lavoro finale previsto per la progettazione delle guide consiste di quattro fasi: acquisizione dell'immagine clinica del ginocchio lesionato del paziente, segmentazione delle immagini per ottenere un modello 3D della regione distale del femore, identificazione della posizione prevista per il tunnel femorale, disegno della guida, con una superficie interna coincidente con quella del femore. Scansioni CT (tomografia computerizzata) sono state utilizzate per il femore sintetico, poiché è necessario un contrasto maggiore. Scansioni RMN isotropiche 3D sono state invece utilizzate per il ginocchio da cadavere, poiché la risoluzione dello standard clinico di imaging (RMN 2D) è risultata insufficiente. Per valutare l'accuratezza delle guide nel tradurre nell'atto chirurgico una posizione del tunnel femorale pianificata nel preoperatorio, nella parete laterale del solco intercondiloideo di ciascuna guida nella posizione definita è stato riportato un foro guida cilindrico di 2 mm di diametro. Sebbene determinare una stima dell'ubicazione dell’inserzione del LCA non fosse tra gli obiettivi di questo progetto, è stato calcolato un valore realistico applicando le linee guida della letteratura. L'apice della cartilagine profonda (ADC) è stato scelto come punto di riferimento anatomico per fornire una stima del centro del fascio antero-mediale (utilizzato come punto di riferimento per la seconda generazione di guide) e del centro dell’intera inserzione del LCA (utilizzato come punto di riferimento per la terza generazione di guide). La progettazione è stata suddivisa in tre diverse fasi, ognuna delle quali ha portato a una generazione di prototipi, seguita da test sperimentali specifici per verificarne i requisiti di progettazione. Gli output di ciascuna sessione di test sono stati utilizzati come input per la fase di progettazione successiva. Gli stampi sono stati creati sul modello di superficie 3D del femore a partire dalle immagini cliniche, per adattarsi perfettamente alla superficie del condilo laterale. Sono state esplorate diverse caratteristiche anatomiche ossee e cartilaginee per garantire un'adeguata stabilità dei dispositivi una volta posizionati. Metodi di analisi Dopo la produzione, ogni generazione di guide è stata testata e valutata in base alla fase di sviluppo e ai requisiti di progettazione. La prima generazione è stata testata sul femore sintetico, per quanto riguarda due caratteristiche: stabilità e congruenza della superficie interna con quella del femore. Il primo requisito è stato verificato tramite ispezione manuale. Per il secondo è stato utilizzato uno scanner 3D a luce strutturata per quantificare eventuali disallineamenti tra le superfici dell’osso e delle guide, che avrebbero potuto portare a problemi di combaciamento e per identificarne l’origine lungo il processo di progettazione. La seconda e la terza generazione sono state invece testate nello scenario finale: le guide sono state impiegate durante un intervento chirurgico di ricostruzione del LCA eseguito su un ginocchio umano da donatore. Le prestazioni della terza generazione di guide sono state giudicate da una valutazione di carattere qualitativo. Le principali caratteristiche funzionali delle guide sono state espresse da tre attributi (posizionamento, stabilità e facilità di inserimento) e ai due chirurghi è stato chiesto di assegnare per ogni attributo un punteggio da 1 a 5 (dal più basso al più alto) per ciascun dispositivo. Nel secondo test sul cadavere, la precisione delle ultime due generazioni di guide è stata valutata con un tracker ottico 3D sul femore dissezionato. Risultati Diviso in tre generazioni, sono stati sviluppati un totale di 10 prototipi. Poliammide 12 è stato scelto come materiale di produzione a causa della sua biocompatibilità certificata e delle buone proprietà meccaniche. Prima generazione: valutazione sul femore sintetico e scansione di superficie I tre prototipi sono stati progettati per ottenere l'adattamento al femore utilizzando il profilo mediale della cartilagine del condilo laterale come vincolo cinematico primario. La copertura cartilaginea, ovvero l'estensione delle guide sulla cartilagine del condilo laterale, è stato il principale parametro variabile oggetto di indagine. La misura intermedia è risultata soddisfacente in termini di dimensione (due delle sue dimensioni non superava i 2 cm), stabilità e facilità di posizionamento, soddisfacendo i requisiti di progettazione e rendendola il miglior prototipo per ulteriori sviluppi. L'errore di fabbricazione è stato generalmente paragonabile alla tolleranza di 0,1 mm fornita dal produttore. Seconda generazione e primo test su cadavere Poiché il test sull’osso sintetico ha mostrato che l’idea impiegata garantisce una buona stabilità e un’appropriata congruenza delle superfici tra osso e dispositivo, sono stati in seguito creati altri quattro nuove guide basate sul modello di superficie 3D del femore da cadavere seguendo lo stesso concetto. Oltre al profilo mediale della cartilagine del condilo laterale, sono state esplorate altre due strutture anatomiche nel femore distale per un ulteriore vincolo cinematico negli altri tre prototipi: la parte più anteriore del profilo mediale della cartilagine del condilo laterale e l'apice della cartilagine profonda (ADC). Anche il meccanismo di manipolazione e inserimento è stato ulteriormente sviluppato per adattarsi al reale scenario chirurgico. Una maniglia indipendente è stata fabbricata con le dimensioni di un tipico strumento artroscopico e un meccanismo a cono quadrato è stato progettato per l'accoppiamento con le guide. È stato possibile posizionare tutte le guide nella posizione prevista senza difficoltà. Una volta posizionate, ciascuna guida ha dimostrato un adeguato grado di stabilità, sebbene le due guide che facevano affidamento sul bordo della cartilagine del tetto della tacca mostrassero un maggiore vincolo alle rotazioni attorno al bordo della cartilagine. Il meccanismo di inserimento e gestione si è dimostrato troppo flessibile ed è stato quindi valutato come inefficace nel facilitare un posizionamento preciso. Terza generazione e secondo test su cadavere Sebbene i modelli precedenti fornissero una stabilità adeguata, l'incompatibilità con la procedura artroscopica ha richiesto una modifica sostanziale del progetto verso un dispositivo più sottile. Sono stati sviluppati tre nuovi concetti per ulteriormente limitare due delle tre dimensioni delle guide a quella dei portali artroscopici. Le nuove guide sono state dotate di una maniglia cilindrica integrata più rigida sia per l'inserimento che per il posizionamento dal portale antero-laterale. Ogni guida ha potuto essere facilmente inserita attraverso il portale antero-laterale, in particolare il più sottile "ADC 2". Il posizionamento nella posizione prevista è stato raggiunto per tutte le guide con supporto endoscopico visivo e facilitato dalla visualizzazione dei modelli disegnati al computer. Nel complesso, la guida “ADC 2” ha ricevuto il punteggio più alto, grazie alla sua facilità di inserimento e posizionamento. Tutte le guide hanno raggiunto una precisione superiore a 5 mm. All'interno della seconda generazione è stata osservata una sostanziale differenza di accuratezza tra i modelli che utilizzavano la parte più anteriore del profilo della cartilagine e quelli che non lo utilizzavano. La guida con il livello più alto di vincolo ha raggiunto la massima precisione e quella con il vincolo più basso ha ottenuto la precisione più bassa. Lo stesso è stato osservato anche per le due guide della terza generazione che avevano un design simile. In media, la terza generazione ha superato la precedente in quanto a performance. Discussione L'ultima generazione di prototipi completamente funzionali è stata progettata dalle immagini 3D RMN di un ginocchio umano da donatore e testata in condizioni chirurgiche standard. Le guide sono state prodotte con un polimero biocompatibile e sterilizzabile, hanno mostrato un'eccellente compatibilità con la procedura artroscopica e un buon adattamento all'anatomia del femore. Le guide hanno ottenuto una precisione più elevata rispetto a un chirurgo esperto, indicata in letteratura come 3-12 mm. Grazie all'utilizzo della risonanza magnetica 3D, questo metodo per il posizionamento accurato del tunnel femorale evita l'esposizione alle radiazioni ed elimina la necessità di tecniche costose, complesse e che richiedono un maggiore tempo per la chirurgia. L’iter di progettazione seguito ha permesso di ottenere delle superfici soddisfacentemente congruenti dalla segmentazione delle immagini cliniche, sebbene sia difficile dire se abbiano influenzato la stabilità dei dispositivi. La pratica standard di imaging non è stata significativamente modificata. Prima di raggiungere l'applicazione clinica, future ricerche dovrebbero concentrarsi sull'ulteriore dimostrazione della robustezza del concetto funzionale superando i limiti metodologici di questo studio, ovvero la dimensione del campione e l’indipendenza delle misurazioni da parte degli utenti. Conclusioni e rilevanza clinica I risultati di questo studio mostrano che le guide femorali prodotte con stampate 3D hanno il potenziale per migliorare i risultati clinici della chirurgia di ricostruzione del legamento crociato anteriore. Se accoppiati con un modo efficace per identificare l'impronta anatomica nativa, i dispositivi sviluppati potrebbero servire come strumento utile per i chirurghi (soprattutto per quelli meno esperti) per migliorare la lassità del ginocchio e l'esito clinico a lungo termine dell’intervento, come la riduzione dell'incidenza dell'osteoartrite.
Personalized intraoperative tools for anterior cruciate ligament reconstruction. Development and evaluation of a 3D-printed patient-specific surgical guide for accurate femoral tunnel placement to improve clinical results in knee surgery
BERJANO ANDOLFI, PABLO
2017/2018
Abstract
Summary Background and rationale The anterior cruciate ligament (ACL) is a ligament of the knee joint that has a crucial role in determining its antero-posterior and rotational mechanical stability. The ACL originates at the posterior part of the lateral wall of the femoral intercondylar notch and descends until its tibial insertion, nearly in the middle of the tibial plateau. From a functional point of view, two families of fibres can be identified in it: the anteromedial (AM) and the posterolateral (PL) bundles. Each of them has a different elongation pattern during knee flexion: the AMB stretches in flexion, while the PLB in extension. The ACL is the most injured ligament of the human body and injuries occur mainly in young and sports-active subjects. The incidence of ACL tears is estimated to be between 1 and 10 in 1000 people per year, with the total number being between 100.000 and 200.000 in the United States. Due to its morphology, ACL tears are mainly caused by a high valgus stress to the extended knee and secondary by hyperextension or extreme internal rotation of the tibia. Given the fundamental role of the ACL in the biomechanics of the knee, its rupture leads to general mechanical instability and abnormal kinematics of the joint, which speeds up the osteoarthritis (OA) progression. The goal standard treatment for ACL tear is the surgical arthroscopic reconstruction (ACLR), whose goal is to restore the pre-injury kinematics of the knee. The ligament is substituted with a graft, typically harvested from the patient's hamstring or patellar tendon, which is fixated in tunnels drilled into the femur and the tibia. The efficacy of the surgical act in restoring the kinematics of the knee has been proven to be highly sensitive to the location where the tunnels are placed and the native footprint (namely, the anatomical point) has been described as the optimal position. However, the original insertion sites are difficult to be determined during the arthroscopical intervention. Indeed, 25-88% of ACL reconstruction failures have been attributed to femoral tunnel misplacement. Identification of the ACL femoral footprint from endoscopic visualization is often not possible and the estimation of its location on the basis of average anatomical guidelines does not give accurate results because of the morphological differences between individuals. Furthermore, accurate execution of the pre-operative plan has been shown to be difficult to achieve even for experienced surgeons, with reported errors between 3 and 12 mm. Computer aided surgery has allowed to reach higher levels of accuracy in tunnel placement with respect to free-hand drilling. However, this technique has seen limited diffusion, since it is still costly in terms of equipment and time. Patient-specific instrumentation has been designed and successfully used in other orthopaedics branches by coupling 3D modelling from clinical images and 3D printing rapid prototyping. Personalized guides seemed to be an inexpensive method to address an issue that is so strongly related to the subject’s anatomical characteristics. We hypothesized that accurate femoral tunnel placement could be achieved with a patient-specific surgical guide. The goal of this thesis was to perform a proof of concept of customized 3D printed molds as a method for accurate femoral tunnel placement that would adapt to the current clinical and surgical practice without further significant costs. A guide that could uniquely fit the patient's anatomy would be an ideal intraoperative tool to translate the preoperative plan (i.e. any concept for femoral tunnel placement) into the actual surgery. This will allow patients to receive truly personalized (anatomical, ideally) ACL reconstructions, regardless the surgeon's expertise, and decrease the incidence of graft failure, thanks to a correct tunnel placement. If coupled with an appropriate method for preoperative identification (or precise estimate) of the native femoral footprint, this will finally allow accurate anatomical tunnel placement, even for all those surgeons that perform only a small number of ACLR every year, since 70% of the ACLR are performed by orthopaedic surgeons that have an experience as low as 20 reconstructive procedures per year). Our research question was: are 3D printed patient-specific surgical guides a feasible solution to reach accurate femoral tunnel placement in anterior cruciate ligament reconstruction? Materials and methods Different designs of 3D-printed patient-specific guides for transferring a templated femoral tunnel position into surgery were developed and evaluated following the standard engineering design process. Overall, three phases of prototype development and evaluation were carried out, each of them addressing the verification of specific design requirements. Design requirements The goal was to develop a device that would uniquely fit the surface of the patient's femur and whose main characteristic would be to accurately report into the surgery the position of the femoral tunnel centre as pre-operatively planned. Different design requirements were set for the device to answer the research question. First, the device should be compatible with arthroscopic surgery and be therefore insertable through a 20 x 10 mm incision. Secondly, the guide should be superior to free-hand tunnel positioning in terms of accuracy: 5 mm of maximum error was set as a reference value. In view of a possible commercialization, the developed guide should be sterilisable and comply with the regulatory standards to certify its biological safety for both patients and users. To adapt as much as possible to the current clinical practice, the design of the guide should, ideally, make use of the standard clinical imaging for ACLR, being 2D MRI. Two characteristics will be then needed by the guide to comply with its final use: fitting, namely perfect congruence to the surface of the femur, and stability, to avoid any slipping on the bone during manipulation and drilling. This could be achieved by selecting an appropriate combination of bony and cartilaginous features to constrain the motion of the guide. Development workflow Three generations of guides were developed with similar procedures. The first preliminary concept was designed on a commercial sawbone and differed from the definitive workflow in terms of imaging technique. The second and third generations more closely reproduced the clinical scenario and were designed for a (single) human cadaver knee. The intended final workflow for the design of the guides consisted in four steps: clinical image acquisition of the patient’s injured knee, segmentation of the images to obtain a 3D model of the distal femur, identification of the intended location for the femoral tunnel, design of the mold with the internal surface coincident to that of the femur for perfect fit. CT (computed tomography) scans were used for the sawbone, since higher contrast is needed for a synthetic femur. 3D isotropic MRI scans were used for the cadaver knee, as the resolution of the 2D MRI imaging clinical standard was found to be insufficient. To evaluate the accuracy of the guides in translating into the surgical act a pre-operatively planned femoral tunnel position, a 2 mm cylindrical guide-hole was reported in the lateral wall of the intercondylar notch of each guide in the defined position. Although determining an estimate of the actual location of the ACL footprint was not among the goals of this project, a realistic value was calculated by applying guidelines from the literature. The apex of the deep cartilage (ADC) was chosen as anatomical landmark to give an estimate of the centre of the AMB (used as a reference point for the second generation of guides) and of the ACL footprint (used as a reference point for the third generation of guides). Three different design phases were carried out, each of them resulted in one generation of prototypes followed by specific experimental tests to evaluate adherence with design requirements. The outputs of each test session were used as inputs for the following design phase. The moulds were created on the 3D surface model of the femur to perfectly fit the surface of the condyles as acquired and segmented from the clinical images. Different bony and cartilaginous anatomical features were explored to guarantee adequate stability of the devices once positioned. Analysis methods After manufacturing, each generation of guides was tested and graded according to its development stage and design requirements. The first generation was evaluated on the sawbone, with regards to two characteristics: stability and fitting. The first requirement was verified through manual inspection. Then, a structured light 3D scanner was used to quantify possible mismatches between the surfaces of sawbone and guides that would have led to fitting issues, and to identify their origin along the design process. The second and third generations were instead tested in the final scenario: the guides were employed during an ACLR surgery performed in a cadaveric knee. The qualitative performances of the third generation of guides was judged by a design attribute assessment. The main functional characteristics of the guides were expressed by three attributes (positioning, stability and ease of insertion) and the two surgeons were asked to assign to each of them a score from 1 to 5 (from lowest to highest). In the second cadaver test the accuracy of the last two generations of guides was evaluated on the dissected femur with a 3D optical tracker. Results Divided into three generations, a total of 10 prototypes were developed. Polyamide 12 was chosen as manufacturing material because of its certified biocompatibility and good mechanical properties. First generation: sawbone evaluation and surface scan Three prototypes were designed to achieve fitting to the sawbone using the medial cartilage edge of the lateral condyle as primary kinematic constraint. Cartilage coverage, namely the extent of the guides on the cartilage of the lateral condyle, was the main variable parameter being investigated. The middle guide coupled intermediate size (two of its dimensions did not exceed 2 cm), satisfactory stability and good ease of positioning, closely meeting the design requirements and making it the best prototype for further development. The manufacturing error was generally comparable to the tolerance of 0.1 mm provided by the manufacturer. Second generation and first cadaver test As the sawbone test showed that good stability was guaranteed by the employed concept and the fitting reached was appropriate, four new molds were created based on the 3D surface model of the cadaveric femur following the same concept. In addition to the medial cartilage edge of the lateral condyle, two other anatomical structures in the distal femur were explored for additional kinematic constraint in the other three prototypes: the most anterior part of the medial cartilage edge of the lateral condyle (the roof of the notch) and the apex of the deep cartilage (ADC). The handling and insertion mechanism was also further developed to suit the real surgical scenario of the cadaver evaluation. A standalone handle was manufactured with the dimensions of a typical arthroscopical instrument and a squared-taper mechanism was designed for the coupling with the guides. All of the guides could be straight-forward positioned in the intended position without hesitation. Once placed, each guide proved an adequate degree of stability, although the two guides that relied on the cartilage edge of the roof of the notch showed a higher constraint to rotations around the cartilage edge. The insertion and handling mechanism proved too flexible and was therefore evaluated as ineffective in facilitating a precise positioning. Third generation and second cadaver test Although the previous models provided adequate stability, the incompatibility with the arthroscopic procedure required a substantial design modification towards a thinner device. Three new concepts were developed to more strictly limit two of the three dimensions of the guides to that of the portals. The new guides were provided with a stiffer integrated cylindrical handle for both insertion and positioning from the antero-lateral portal. Every guide could be easily inserted through the antero-lateral portal, especially the thinner, “ADC 2” model. Positioning in the intended position was reached for all the guides with visual endoscopic support and facilitated by visualization of the computer-drawn models. Overall, the ADC 2 guide received the highest score, thanks to its ease of insertion and positioning. All the guides reached an accuracy higher than 5 mm. Within the second generation, a substantial difference in accuracy was observed between the models that used the most anterior part of the cartilage edge and the ones that did not. The guide with highest level of constraint reached the highest accuracy and the one with lowest constraint achieved the lowest accuracy. The same was valid also for the two guides of the third generation that had similar design. On average, the third generation outperformed the previous one. Discussion The last generation of fully functional prototypes was designed from the 3D MRI images of a cadaver knee and tested under standard surgical conditions. The guides were manufactured with in a biocompatible and sterilisable polymer, showed excellent compatibility with the arthroscopic setup and good fitting on the anatomy of the femur. The guides achieved higher accuracy than an average experienced surgeon, indicated in the literature as 3-12 mm. Thanks to the use of 3D MRI, this method for accurate femoral tunnel placement obviates radiation exposure and eliminates the need for expensive, complex and time-consuming techniques. The followed development workflow reached satisfactory fitting from the segmentation of clinical images, although it is difficult to say whether it influenced stability. Imaging standard practice was not significantly altered. Before reaching the clinical application, future research should be focused on further demonstrating the robustness of the functional concept by overcoming the methodological limitations of this study, namely sample size and inter-user reliability of measurements. Conclusions and clinical relevance The results of this study show that 3D printed femoral guides have the potential to improve the clinical outcomes of ACLR. If coupled with an effective way to identify the native anatomical footprint, the developed devices could serve as a useful tool for surgeons (especially the most unexperienced ones) to improve knee laxity and long-term clinical outcome of ACLR, like reduction of osteoarthritis incidence.| File | Dimensione | Formato | |
|---|---|---|---|
|
Development and evaluation of a 3D-printed patient-specific surgical guide for accurate femoral tunnel placement in ACLR.pdf
non accessibile
Descrizione: Thesis_final manuscript
Dimensione
14.21 MB
Formato
Adobe PDF
|
14.21 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/142678