Thybout M. Moojen Carpal Kinematics

Summary and Discussion
By definition, the carpus is considered unstable if it shows a symptomatic dysfunction, is not able to bear loads, and does not exhibit normal kinematics during any portion of its arc of motion [81]. Based on this dentition, the term carpal instability should not be used to designate a specific type of pathology but to define a syndrome characterized by the loss of the normal relationship between the articulating bones, resulting in abnormal motion (dyskinematics), or altered transfer of load (dyskinetics) [82]. Carpal instability comprises a spectrum of severity, ranging from very subtle (and until now undetectable) ligament lesions, to static deformities. It is possible that carpal instability is a progressive phenomenon, in which (partial) ligament lesions gradually and sequentially lead to the stretching and rupture of other ligaments. The availability of a dedicated screening method for ligamentous injuries of the carpus is highly desirable, because early detection of these injuries, if present, is important for preventing late problems that cannot be solved without residual functional limitations.
Only very recently we entered an era in which sophisticated techniques have been developed that make it possible to accurately quantify 3D carpal kinematics in vivo [25.29, 83]. The method described in this thesis is one of these techniques. It may have the potential to help answer questions, such as how carpal instability progresses and what the (long-term) effects of surgical interventions may be. Animations of the carpal bones during wrist motion help to understand the complex 3D carpal movements and contribute to the recognition of pathological wrist motion and the development of implants and operative techniques. Finally, data and animations can easily be shared with and used by others by means of the internet. This has been suggested by the Editorial of January 2003 of the EditorinChief of the Journal of Hand Surgery (American Volumes): .The online version of the Journal expands the traditional format of the Journal. It allows unlimited colour illustrations as well as the presentation of video clips, as noted in this issues article by Moojen and colleagues. [84]
One of the basic procedures in this 3D analysis is the matching of the individual carpal bones in the different postures. For this purpose several authors have used cortical margins to define the carpal bones [27]. Because we used the complete interior 3D information of the carpal bones, i.e. the very specific trabecular structure, translations and rotations can be measured much more accurately than with algorithms based on the surface information of the bones only. This is very important, since the finite helical axes parameters, which are used in this thesis to describe the motion of the bones, are very sensitive to errors in rotation angles, especially when small increments in motion are applied [46]. No repeated segmentation or manual landmark detection is required in our technique, which dramatically reduces the amount of human operator intervention that is required. The matching technique is fully automatic, in contrast to the techniques used by other authors.
The methods used by Wolfe et al. and Sun et al. were limited to the analyses of in-plane motions of the carpal bones during flexion-extension motion of the hand. Outofplane motions could not be quantified, because these methods used 2D sagittal CT scans [28,29]. As a consequence, scan sessions had to be discarded if the shapes of the bones varied signicantly from image to im age. With our method it is possible to measure both inplane and outofplane motions of the carpal bones during flexion-extension motion (FEM) and ulnar-radial deviation (RUD) of the hand because of the use of 3D CT scans. In our study we used spiral CT scanning, in contrast to the conventional (sequential) CT techniques used in most of the earlier studies. In spiral CT the body is moved continuously through the revolving gantry instead of being shifted stepwise, which is the case in conventional CT scanning. The main advantage of the spiral CT technique is that the images are obtained at higher speed, and can be reconstructed at smaller increments (interslice distance), which is especially useful for high-resolution 3D imaging. In order to validate our method, we first performed an accuracy analysis by scanning cadaver hands in different positions. This study showed the mean error in rotation to be less than 0.4 fl and in translation less than 0.5 mm [66].
In the five chapters of this thesis that have been published as articles, measurements have been reported of the translations and rotations of the carpal bones during different movements (radial-ulnar deviation and flexion-extension) of the hand. The description of what exactly was reported in these chapters was very short. Also the precise definition of the translation parameters in chapters 3 to 6 was not presented. Therefore a short overview is given here of the conventions that have been used and the parameters that have been reported.
In this thesis the finite helical axis (FHA) method has been used, as it is suitable to express the rotation and translation parameters in 3D joint kinematics. In this method the joint movement parameters are expressed in a consistent way [51] as opposed to the Euler and Cardan methods [85.87], where the sequence of the component rotations about the coordinate axes need to be specified. The FHA method is based on the principle that any displacement of a rigid body in space can be described as a translation along a particular axis (the helical axis) and a rotation around this axis. Woltring [51] introduced an attitude vector by multiplying the unit vector .n in the direction of the helical axis by the helical rotation angle theta (see section 2.3). For the atti tude vector the helical axis is used that describes the motion from a reference position and orientation to the current position and orientation. In this the sis, this attitude vector has been used to quantify the rotations of the carpal bones. In chapters 2.6 the projections of the attitude vector onto its respective joint axes, the radialulnar deviation axis (zdirection), the flexionextension axis (xdirection), and the supinationpronation axis (ydirection) have been reported.
For the reporting of the translations of the carpal bones, two different conventions have been used. In chapter 2 the translation along the helical axis has been reported for the capitate of one volunteer in radial-ulnar deviation and flexion-extension motion, whereby the axis was calculated for the motion between the neutral reference and the other positions. In chapters 3.6, this parameter has not been reported. Instead, in order to facilitate the interpretation of the kinematic data and the comparison of these data with the literature, the authors presented data on the translations of the carpal bones relative to the neutral position in the three coordinate directions. These translations were calculated by taking the distance of the centroid of each carpal bone between the neutral position and all other positions at the different angular displacements of the hand. For the calculation of the centroids an area weighted aver age was used for all triangles of the triangulated surfaces of the carpal bones. We compared our in vivo results with results derived in the past by the most accurate in vitro studies. Our data largely concurred with data presented in the literature; we found, as was also described by Craigen and Stanley, Nuttall et al., GarciaElias et al., and Ferris et al., a spectrum of movements of the bones of the proximal carpal row compared to the distal carpal row, particularly during RUD of the wrist [71.73, 88]. In some wrists the proximal row predominantly flexed and in others it mainly translated in a radial-ulnar direction. The intercarpal motions among the bones of the proximal car pal bones, however, showed very uniform and predictable kinematic patterns with smaller variations than between the proximal and distal carpal rows. The motion patterns may be induced by a combination of ligamentous laxity, the shape of the articular surfaces and the constraining effects of the musculotendinuous structures crossing the joint, particularly between the proximal and distal carpal row. It might be hypothesized that the predictable and uni form motion patterns within the proximal carpal row are due to the tight and short intrinsic carpal ligaments. The larger spectrum of movements between the proximal and distal carpal row during RUD of the hand might be a result from a combination of the bony anatomy and the relative laxity of the longer extrinsic ligaments. As GarciaElias described, lax individuals have scaphoids that appear to be less tightly bound to the distal carpal row, allowing greater scaphoid (and thus also of the lunate and triquetrum) rotation in the sagittal plane and thus requiring less deviation in the coronal plane, to achieve maximal wrist RUD. The opposite is true for the tighter individuals. The whole proximal row must deviate more laterally to compensate for the lack of scaphoid flexion-extension. GarciaElias hypothesized that the less constrained the scaphoid is in the sagittal plane, the higher the incidence of periscaphoid ligamentous injuries will be. This may be one reason why lax wrists are so frequently symptomatic if overloaded, because they have relatively more vulnerable carpal kinematics [72].
At present, there is no true real-time (i.e., for example, 25 frames per second) 3D imaging of the wrist available. Instead, all current 3D imaging techniques consist of the acquisition of a number of static 3D images of the hand in a number of postures. This is a clear drawback of these methods, because the influence of the forces of the tendons crossing the joint and the position of the hand might be different during motion, compared to repetitive positioning. Imposed by the spectrum of kinematics between individuals, the very small motions, differences in forces, shapes and ligaments, the registration method must be very specific. Because accurate quantification of 3D continuous motion is not possible yet, we studied repetitive positioning of the hand. It is imminent that the motion steps must be as small as possible, and the forces crossing the joint as standardized as possible.
Until now, the number of hand positions analyzed in the studies known in literature is limited due to radiation safety concerns [25.27]. Due to the inherent high accuracy of our method, it was possible to use a lowdose CT technique in the present study in order to reduce the radiation exposure for the CT scan of each posture. Therefore, the hand could be imaged in a large number of postures, representing a number of small increments during motion without excessive exposure to ionizing radiation and without a significant decrease in accuracy. As a consequence, the effective dose of one complete investigation of a wrist is very low, and is in the order of the effective dose of one or two chest xrays.
The relative slow capture time for imaging technology currently available on standard machines requires that the subject be stationary while the scan is conducted. The musculoskeletal control necessary to perform such fixed (static) positions is different from that used in a continuously moving action, and the kinematics results may also differ. In hand pathologies, the relation ship between motion, pain, and load in the wrist are of interest, but hard to quantify. The role of axial compression, working across the joint due to the tone of the forearm muscles, is emphasized through reports where in vitro studies experimentally produced various carpal malalignment or instability patterns by sequentially cutting the various ligaments, and failed in their objectives unless the experiment was combined with an axial compressive load [89, 90]. Axial load has been used to identify dynamic instability patterns by precipitating subtle carpal malalignment through clenched fist radiographs [91]. Various factors that contribute to the axial load across the joint are the tone of the skeletal muscles, viscoelastic properties of the soft tissues includ ing the ligaments of the wrist and the gravitational force, which depends on the position of the hand. Few in vivo kinematic studies of the wrist have been conducted with the joint loaded and none has compared kinematics at different load levels.
Kobayashi et al. studied 13 cadaver wrists under axial compression. They found the complete proximal row to flex, radial deviate and supinate during 98 N axial compression in the neutral position of the hand, without a signi fi cant change in the intercarpal relation [1]. ValeroCuevas et al. performed an in vivo study in which they studied wrist motion while generating torques of zero, 1.1 and 2.2 Nm in a planar, unidirectional flexion motion. They found that the wrist did not behave like a smooth mechanism when generating torque, and also that the differences were not significantly different between subjects at a given load [92]. They postulated that the joint forces which accompany the generation of wrist torque induce more irregular carpal motion. However, with the method they used in their study, carpal kinematics could not be analyzed. Gupta studied the role of physiological axial loading on the carpus [93]. Lateral plain wrist radiographs of uninjured wrists were made in 20 patients before and after giving general anaesthesia along with muscle relaxants, and were repeated after applying traction in line with the long axis of the radius. They found that anaesthesia caused the scaphoid and lunate to extend with no change in scapholunate angle, while traction caused the scaphoid (and lunate) to rotate further dorsally. From the above mentioned studies it might be concluded that skeletal muscles under physiological conditions exert some flexion torque over the proximal carpal row without a significant change of the intercarpal relation, as had already been concluded by Kobayashi et al.
We decided to analyze wrists under resting tone loading. We found our data to correspond very well with the in vitro data derived from the literature, although these studies used very different wristloading protocols. In a future study, the wrists of volunteers should be scanned with resting tone loading, and with, e.g., 50 N loading (using the Jamar dynamometer attached to the CT table), 100 N, and 200 N, and look for a change in kinematics. It may be very well possible that lax individuals tend towards a more translational kinematic pattern of the proximal carpal bones because of the decreased influence of extrinsic ligaments during compression of the wrist, and because of the forced 92 Carpal Kinematics extra flexion of the proximal row. It may also be possible that rigid individuals behave the same under extra loading because of their relative tight extrinsic ligaments which not allow more flexion of the scaphoid and the rest of the proximal row. These changes will be most pronounced during RUD of the hand, because of the larger inter individual kinematic spectrum during RUD compared to FEM.
It would also be very interesting to study carpal kinematics when the hand is moved in a restrained fashion from extension to flexion and from radial to ulnar deviation. This is possible by using the apparatus described by Gupta; a custommoulded handle for each subject connected to a chair. In our study environment it can be attached to the CT table. A suspended weight produced a torque via a circular cam. They studied torques of zero, 1.1, and 2.2 Nm throughout the range of motion [93].
In our series, only the right hands were scanned to exclude the possible intraindividual variations. If both wrists of the same person would show the same kinematical patterns, patients could be used as their own reference. Wolfe et al. recently studied in vivo both wrists of 10 uninjured volunteers of which 5 women and 5 men. CT scans were obtained in the neutral position of the wrist and in 4 different positions of wrist flexion and extension. Their study showed no significant kinematic differences between the right and left wrists. In addition, they were unable to show a difference between the rotation, translation, and orientation of male and female subject kinematics. Similarly, there did not appear to be a difference between dominant and nondominant wrists. Their study demonstrated that in vivo carpal motion during flexion and extension between subjects is strikingly similar [28]. These results con firm our data of in plane carpal motion during flexion and extension of the hand. We found, however, large interindividual variations in carpal kinematics during radial and ulnar deviation of the wrist between the proximal and distal carpal row, although the intercarpal motions within the rows are very constant. From the above mentioned study of Wolfe et al. one can not conclude that motion patterns of men and women are the same, neither that the contralateral hand can be used as a reference, because only FEM was studied. Recently Crisco et al. published a very interesting study in which they analyzed patients with arthroscopically veri fied scapholunate ligament lesions, and compared the kinematics of the injured wrists with data of the uninjured, contralateral wrists, and also with an existing database of healthy male and female volunteers [27]. Their results con firm again that SL ligament tears alter carpal bone kinematics. They found the lunate to extend 27 fl more in the neutral posture of the wrist in the injured wrist, when compared to the con tra lateral side. This difference can easily be detected with our method, with standard deviations of maximally 3 fl. It may be that there is a population of individuals predisposed to ligament injury due to abnormal carpal kinematics. Other authors have shown bilateral soft tissue defects after unilateral wrist injury in 60% to 100% of the patients [94, 95]. In patients with a range of in juries, Feipel et al. found also no significant kinematic differences between the injured and the contralateral, asymptomatic wrist [96]. There was although a significant difference between both wrists of the injured patients and the wrists of the uninjured volunteers. Further research is needed to con firm and understand the finding of bilateral abnormal carpal bone posture and kinematics in subjects with unilateral wrist ligament injuries.
For diagnostic purposes, a selection of the wrist postures we analyzed in our studies could be chosen. With 60 fl of wrist flexion, e.g., the scaphoid flexes 19 fl more than the lunate, with a standard deviation of only 3 fl, which implies a predictable and uniform motion pattern. Most likely, the difference between scaphoid and lunate flexion will increase more than those 3 fl when the SL ligament is torn completely. Kobayashi con firmed in cadaver wrists that, when a complete SL dissociation is reproduced in cadaver wrists, the constraining effect of the lunate on the scaphoid disappears, and the intercarpal rotation difference between the scaphoid and lunate will probably increase. When interested in diagnosting lunotriquetral (LT) ligament lesions, it would be possible to scan the injured wrist in, e.g. 60 fl of extension or 20 fl ulnar deviation. When the LT ligament is torn, it is very probable that the extension of the triquetrum increases, forced by it connection to the hamate, while the lunate extends less. Ritt et al. performed a study in which they studied the effect of sequential sectioning the ligaments of the lunotriquetral joint in cadavers [60]. They found that the intercarpal motion between the triquetrum and lunate changed significantly after sectioning the palmar and proximal region of the lunotriquetral ligament. Gross malalignment became apparent only after additional manipulation (1.000 cycles of cyclic load) of the specimen. With our method it should even be possible to detect a partial LT ligament lesion before the repetitive loading of the wrist occurs, i.e. in an acute situation. Ritt et al. found, after partially sectioning the LT ligament, that the lunate extended 12.6 fl more than the triquetrum, with the hand in 30 fl ulnar deviation. The intercarpal data with the intact ligament are exactly the same as our data, i.e. an intercarpal difference of 4 fl. The standard deviation in our study in this situation is 2 fl, and the maximal error is 0.4 fl. From these data it might be concluded that these differences could be detected.
We studied the wrists of volunteers in FEM and RUD to maintain repro ducibility in the experimental protocol. Measurements of motion in these anatomically de fined planes will not generate a complete kinematic description of the normal wrist, which is more complex and involves multiplanar motion. The deviation of the wrist in any direction describes an oval plane. Outofplane motions as described for the carpal bones are actually a consequence of an arbitrarily imposed orthogonal plane of measurement and shows that natural wrist motion occurs in a unique plane with smooth uniplanar motions of the individual carpal bones. Analysis of wrist kinematics should be done relative to the functional motion of the wrist and not to the plane of the metacarpals.
Future perspectives
All future treatment methods will benefit from improved imaging methods that will show soft tissue or bone damage from every position and in all functional modes before, during and after treatment. Each wristtherapeutic procedure has so many variables that there is a need for an interactive learning process, in which questions can be asked, anatomy displayed, and alternative methods tried. New procedures will be simulated and their likely effects on the kinematics of the system observed well before they are actually performed on the patient. By combining measured 3D kinematic data with predictions from a biomechanical model of the wrist, the diagnosis of injuries can be improved by revealing the exact mechanism that resulted in the observed kinematic abnormality [97]. The data of normal wrists can be used as input data for the development of kinematic models of the normal wrist for diagnostic purposes. Such models may be used for automatic detection of pathological motion patterns. The current software implementation is aimed at statistical analysis of carpal kinematics. To employ this software in clinical practice, a user-friendly interface should be developed.
In the future, the introduction of a combination of 3D CT with fluoroscopy may be expected. This has been proven to be a fruitful approach to obtain real time 3D kinematics of the knee joint [98]. Currently an application for the wrist is being developed to obtain 3D carpal kinematics in vivo (G.J. Streekstra, personal communication). To this end the 3D geometry of the carpal bones as reconstructed from a single CT scan is combined with the dynamic 2D images acquired with fluoroscopy. This would make it possible to study realtime carpal kinematics. Since CT and fluoroscopy of the wrist can be done within a reasonable time span and without exceptional efforts, this approach could make early screening of the injured wrist within reach of daily clinical practice. Problems which will be encountered are segmentation and matching dif ficulties. The cinematography does not provide internal 3D information on the trabecular structure, making the matching procedure less accurate, and the overlapping contours of the carpal bones and narrow joint spaces will pose problems for the segmentation procedure.
We expect that quantitative in vivo 3D studies on carpal kinematics, especially with regard to dynamic wrist motion, may have future diagnostic applications, particularly following ligament injuries that do not present with overt carpal malalignment. In the future these in vivo registration methods could provide valuable information on longterm therapeutic results and possibly predict the outcome of operative interventions.
Samenvatting
DE POLS wordt beschouwd als het meest ingewikkelde gewricht van het menselijke lichaam. Dit komt doordat het gewricht bestaat uit 15 bot stukken (radius, ulna, 8 carpalia en 5 metacarpalia) met veel en onregelmatig gevormde contactvlakken (zie Figuur 8.1). Deze botten zijn nauw met elkaar verbonden door tientallen bandjes, zonder dat deze de grote bewegelijkheid van de pols mogen belemmeren.
De bewegingen van de hand ten opzichte van de onderarm worden mo gelijk gemaakt doordat de acht handwortelbotjes (carpalia) allemaal gecompliceerde bewegingen maken ten opzichte van elkaar, zonder dat deze gestabiliseerd hoeven te worden door volumineuze spieren, zoals het geval is bij het heupgewricht. Dit maakt de pols echter wel relatief kwetsbaar. In de pols is alles van elkaar afhankelijk wat zich uit in het feit dat als ergens in de keten een probleem ontstaat, hoe klein ook, dit vaak gevolgen heeft voor de rest van het gewricht. Dit resulteert vaak uiteindelijk in onherstelbare slijtage. Invaliditeit ten gevolge van polsaandoeningen is aanzienlijk en komt steeds meer voor. Pijn en verlies van bewegelijkheid van de pols zijn direct van invloed op de functie van de hand, en dus van de persoon als geheel. Problemen aan de pols kunnen het gevolg zijn van een ongeval, maar ook bijvoorbeeld van chronische gewrichtsaandoeningen, zoals reuma. Bij veel patiënten is er echter geen duidelijke oorzaak te vinden met de diagnostische mogelijkheden van vandaag (bijvoorbeeld RSI).
Tot zeer recent was er geen methode beschikbaar waarmee de bewegingen (kinematica) van de carpalia bij pati¨enten nauwkeurig gemeten kon worden. Sinds het einde van de negentiende eeuw is men bezig om de carpale kinematica te analyseren. Dit was vooral bedoeld om de behandeling van reuma en posttraumatische afwijkingen te verbeteren. Men bestudeerde toen (2dimensionale) röntgenfoto’s, waarmee men uiteraard alleen de positie van de afzonderlijke botten kon beschrijven (zie Figuur 8.1). In de loop van de vorige eeuw zijn er meerdere methoden ontwikkeld waarmee men de bewegingen van de carpalia wel kon meten, maar deze technieken waren invasief van aard. Deze methoden maakten gebruik van metalen kogeltjes of pinnetjes, geplaatst in de afzonderlijke botten. Hiermee werd het mogelijk om nauw keurig de bewegingen te meten in 3 dimensies, echter alleen bij polsen van overledenen. Het is denkbaar dat de bewegingen van de carpalia in deze polsen anders zijn dan bij levenden, en zeker als er in de carpalia eerst pinnen of kogeltjes zijn geïmplanteerd. Ook is het bij polsen van overledenen uiteraard niet mogelijk om de eventueel veranderde bewegingen van de carpalia te diagnosticeren na een val, of om de lange termijn resultaten van operaties of revalidatie te vervolgen. Deze beperkingen hebben ertoe geleid dat men door is blijven zoeken naar methoden welke wel gebruikt kunnen worden bij patiënten. Dit resulteerde in de laatste jaren van de vorige eeuw in de ontwikkeling van enkele, zij het vooralsnog experimentele, technieken. Hoofdstuk twee van dit proefschrift beschrijft een van deze nieuwe methoden. Bij deze methode wordt gebruik gemaakt van een spiraal CT scanner. De polsen van patiënten en vrijwilligers worden in verschillende standen van de hand gescand. De opname van de hand in de neutrale stand wordt gemaakt met de reguliere stralingsdosis, en de andere standen van de hand in flexie en extensie en radiaal en ulnair deviatie worden gemaakt met een tiende van deze dosis. Hierdoor wordt de totale hoeveelheid straling beperkt. De beelden van de verschillende botten worden zodanig bewerkt (gesegmenteerd en `gematched’), dat de rotaties en translaties in 3 dimensies kunnen worden berekend. Met deze methode blijkt het goed mogelijk om de bewegingen nauwkeurig te meten en te visualiseren.
In het derde hoofdstuk worden de methode en de resultaten die wij verkregen hebben bij gezonde vrijwilligers vergeleken met de verschillende methodes welke in het verleden zijn gebruikt om de carpale kinematica te kwantificeren (in maat en getal uit te drukken). De voor en nadelen van de ver schillende methoden worden besproken en de meetresultaten worden geanalyseerd. Bij deze vergelijking blijkt dat de methode beschreven in dit proef schrift zeer nauwkeurig is en dat de resultaten goed overeenkomen met die van de meest nauwkeurige methoden tot nu toe. De voordelen van de in dit proefschrift beschreven methode zijn meervoudig. Patiënten en vrijwilligers kunnen geanalyseerd worden. De stralenbelasting is beperkt waardoor meer verschillende posities van de pols gescand kunnen worden. Onze methode is nauwkeuriger dan de andere in vivo methoden. Alle acht carpalia kunnen bestudeerd worden. En het belangrijkste voordeel is dat de rotaties en translaties in drie dimensies nauwkeurig gemeten kunnen worden. In het vierde hoofdstuk wordt een complete beschrijving gegeven van de kinematica van alle carpalia. Dit hoofdstuk bevat de eerste database in de literatuur van alle acht carpalia gedurende flexie, extensie, radiaal en ulnair deviatie. We konden de hypothese bevestigen dat er een spectrum aan be wegingen van de proximale carpale rij ten opzichte van de distale carpale rij zou bestaan. De bewegingen tussen de botten van de proximale rij, echter, vormen een vast patroon en zijn daarom goed voorspelbaar. Dit houdt in dat deze gegevens gebruikt zouden kunnen gaan worden om bepaalde ligamentafwijkingen in een vroeg stadium aan te tonen. In de toekomst kan het mogelijk zijn deze database te gebruiken om de gegevens van patiënten met polsproblemen mee te vergelijken.
Hoofdstuk vijf beschrijft de kinematica van het scafoid. Recente studies beschrijven een spectrum van bewegingen van het scafoid gedurende radiaal en ulnair deviatie van de pols. Dit vermoeden konden wij met onze nieuwe en nauwkeurige meetmethode bevestigen. Doordat het scafoid niet bij iedereen hetzelfde beweegt moet de behandeling van een ligamentprobleem of een bepaalde botbreuk ook mogelijk worden aangepast aan het bewegingspatroon van de desbetreffende patiënt. Ook is nu duidelijk dat bepaalde tests die nu voor diagnostiek gebruikt worden niet altijd even betrouwbaar zijn, omdat die uitgaan van een standaard bewegingspatroon.
Hoofdstuk zes bevat de beschrijving van de kinematica van het pisiforme. Dit betreft de eerste studie in de literatuur die dit nauwkeurig beschrijft. Over de functie van dit botje bestaat nog weinig duidelijkheid. Het blijkt een volledig ander bewegingspatroon te hebben dan de andere zeven carpalia, zodat het te billijken valt het pisiforme niet tot de carpalia te rekenen. Concluderend kan gesteld worden dat deze methode ertoe kan bijdragen om polsproblemen in een eerdere fase te ontdekken dan tot nu toe mogelijk is, waardoor, zonodig, vroeger kan worden ingegrepen en onherstelbare schade mogelijk voorkomen kan worden. Ook zal het in de toekomst mogelijk zijn om de resultaten van polsoperaties en andere behandelingen beter te vervolgen en het te verwachten therapeutische effect beter te voorspellen. Met behulp van deze gegevens kan het mogelijk zijn om nieuwe operatietechnieken en implantaten te ontwikkelen. Het visualiseren van de kinematica van de carpalia vanuit elk gewenst gezichtspunt maakt deze ingewikkelde materie beter te begrijpen. In de toekomst wordt het misschien mogelijk om deze methode te combineren met röntgendoorlichting, waardoor het mogelijk zou worden om continue polsbewegingen te analyseren.