Will pleural fluid affect surface wave speed measurements of the lung using lung ultrasound surface wave elastography: experimental and numerical studies on sponge phantom?

1 Will pleu ral flui d affect surface w av e speed measur ement s of the lung using lung ul trasound surface w av e elastograp hy : e xperimental and n umeric al stud ies on sponge phantom ? Boran Zhou 1 , Xiaom ing Zh ang 1 1 Departm ent of R adiolog y, Ma yo Clinic C ollege of Med icine and Sc ience, 200 1st St SW , Roc hester, MN, 55905, U SA. Correspond ence: Xiaom ing Zhang, PhD Zhang.x iaoming @m ayo.edu . Departm ent of R adiolog y, Ma yo Clinic, 200 1 st St SW , Rochest er, MN, 55905, U SA. 2 Abstract P leural eff usion m anifested as com pres sion of pleu ral flui d on the lung pare nch y ma , c ontrib uti ng to hypoxem ia . M edical proc edure s uch as drain age of plural f luid rele ases this compr ession and increase the oxygena tion. H owever, the ef fec t of pleura l ef fus ion o n the elasticit y of lung parenc hyma is unk nown. B y using the lung u ltrasou nd s urfac e wave elastogra ph y (LUSW E) and finite e lem ent method ( FEM) , the eff ect of pleur al ef fus ion on the elast icit y of s uperfic ial lun g parenc h yma in ter m s of s urf ace wave speed meas urem ent was e valuat ed in a sp onge phantom s tudy . D if ferent th ick ness of ultrasou nd tr ans mis sion gel s imulated as pleura l fl uid was insert ed in to a c ondom w hich was plac ed b etween t he s ponge an d standoff pad. A mec hanical shak er was used to g ener ate vibrat ion on the s pon ge phant om at differ ent frequenc ies r an ging fr om 100 to 3 00 Hz while ultras ound tr ansduc er was us ed t o capture t he m otion for meas urem ent of sur fac e wave speed of the s pong e. FEM was cond ucted bas ed on th e e xper im ental setup and num er ically assess the influence of pleur al ef fus ion on the surf ace wa ve speed of the sponge . W e found the influ ence of thic knes s of ultras ound tr ansm iss ion gel was stat istic all y insign ificant on t he surf ace wave sp eed of the sponge at 100 and 150 H z both f rom exper im ents the FEM. Keywords: L ung ultr asou nd surf ace wave e lastogr aph y (L U SW E ) ; p leural eff usion; lung par enc hyma; lung spon ge pha ntom ; finite el em ent m odeling. 3 I ntroduction P leural ef fus ion, freque ntl y encounter ed i n critica lly ill patients hospital ized in inte nsive care un its (ICU) , is th e accum ulation of ex cess fluid in the p leura l cav it y which is th e flui d - filled s pace that sur rounds the lungs. I t i nduces r estr ictive s yndrom e s and inc reas es the i ntrap ulmonar y shunt b y com pressing th e lung parench yma. St ud y has sho wed that the ple ural eff usion m a y contribute to h ypoxem ia under mechanic al ventilati on [1] . I n order to increase oxyge n deliv ery an d ox ygen consum ption , m edica l procedure s such as d rain age of pleural f luid is perfor m ed to increas e functio nal res idual cap acit y and impr ov e ox y genatio n . H em od ynamic and pulm onary p aram eters , such as bl ood pres sure, s ystemic vascular resistanc e, peak airwa y pressure , were collec ted befor e and af ter the f luid was drain ed [2]. H owever, no tec hnique is able t o evalu ate the degre e of restorat ion of the f uncti on of lung p arench yma. I n the ICU, the di agnosis of pleural eff usion relies on the a nteropos terior chest r adiogra ph y obtained a t beds ide [ 3] . H owe ver, it exposes pat ients to a hi gh dose of r adiatio n. P le ural s onograp hy is an alter nati ve im aging m odalit y . It is a highly portable and widel y accept ed diagnos tic technique for identif ying pleur al dis ease [4, 5] . It per m its imaging of pleural ef f usion and other pl eura l pathol og y. I n addition, u ltraso nogra phy is able to guid e thorace ntes is for pleural int erventi ons. Normal viscera l and parietal pleura are app ose d and 0 .2 - 0.3 m m thick. P leural eff usions with par ietal p leural t hick ness > 1 0 mm, and diaphr agm atic thi cknes s > 7 m m predict un der lying m alignanc y with h igh spec ificit y. I t has been shown that a m inim um pleu ral eff usion dept h of 1. 2 cm bet ween the visceral and par ietal ple ura h as be en recom mended to perfor m diagnostic t horacen tesis [6 ]. T horac entesis is us ual ly per form ed to relie ve the co m pression of pleural eff usion on th e lung parench yma. F ew study evaluat ed the restoration of lung parench yma after drain age of pl eural f luid. Function of lun g parenc h yma is heavil y depende nt o n its elastic ity. I n ord er to n oninv asivel y evaluat e t he elasticit y of l ung par ench ym a for patients with in terstit ial lung dis ease (I LD) , w e r ecentl y developed a lun g ultrasound sur face wav e e lastogra ph y ( L U SW E) to m eas ure the sur face wave s peed of lung which is correlate d with lung elas tic ity [7 - 9] . No pleur al ef fusion was o bserv ed for IL D patien ts so u ltrasou nd propagat ion can penetr ate the thor acic m uscle and motion of lu ng s urfac e can b e capt ured b y ultrasoun d imaging. F or the pat ients w ith pleur al effus ion, the e ffec t of pleural f luid on the s urf ace wave speed of 4 lung par ench yma in LU SW E is unknown. T heref ore, t here is a press ing nee d to d evelop a pha ntom model to s ystem aticall y inves tigate t he eff ect of pleural f luid on the m easurem ents of LUSW E. W et foam dr essing m aterial has been used for lung ul trasound sim ulation m odel s to teach nov ice physicians to perfor m lung ultrasou nd i n cl inic al s ituations [10] . An e c onom ical s ponge phant om was used for under standin g and r esearc hing re verber ation artifac ts in lun g ultras ound gi ven it s sim ilar micros tructur e with lung parench y ma [11] . M oreov er, with its a vail abilit y, rele vant phant om models for a system ic s tudy of induc ed diseas e states , such as pu lm onar y edema, c an be gen erated. The objec tive of this stud y was to de velop a phant om m odel f or eval uating the eff ect of pleural fluid on surfac e wave sp eed in L U SW E . U ltras ound trans m ission ge l t o sim ulate p leural fluid was squeezed in to a c ondom that was plac ed b etween the acoustic s tandof f pad a nd sponge phantom . I n LU SW E , a shak er was used t o gener ate a vibrat ion on the sur face of stand off pad , and t he w ave propagat ion on the surfac e of the s ponge p hantom was m easured b y using ultrason ic im aging and analysis. A F EM m odel was deve loped to sim ulate the wave propaga tion in the s ponge phantom according to the experim ental set up and c om pare w ith exper imental m eas urem ents. T he res t of t he pa per is st ructured as f ollows: we de sc ribe the set u p for the sponge phan tom model in the Materia ls and Metho ds s ection; we pr esent results th at eva luate th e eff ects of thick ness of ultrasound t rans m ission gel in t he resu lts sec tion; we fi nalize th e paper with discussion and conc lusio ns. 5 Mat erials and M et hods T he exper im ental setting consist ed of th e follo wi ng parts : (1) house hold s pon ge ( Oc elo uti lit y sponge, 3 M, St. Pau l, MN ); (2) ultras ound transm ission gel ( Aquas onic 10 0, P ark er Laborator ies I nc, Fairf ield , NJ ); (3) an acoust ic st andof f pad ( Aqu aflex ; Park er Laborat ories I nc, F airf ield, NJ) . The ac oustic standoff pad is m ade of a ge l m atrix f ree of air bubbl es elim inating t he air - f illed s pace bet ween the transducer and th e spo nge phant om . W ithout a standof f pad, ex tra - th oracic tiss ues will not be satis fact orily imaged. T he f luid com ponent of a pleural ef fusion m a y have echogenit y which is character istic of the pres ence of c ellular ity. A ir bu bbles within ple ural flu id, which m a y occur with esophagea l - pleur al fis tula or a gas - form ing infec tion will exhibi t multi ple m obil e echogen ic f oci within pleural f luid th at re present air bub bles [12] . U ltrasou nd tra nsm iss ion gel has the sim ilar ec hoge nity as pleural f luid and can cont ain a ir bub bles i n it. Spong e has been sho wn to have sim ilar m icr ostructure as lung par ench yma. U ltras ou nd tra nsm ission ge l w as sq ueezed into a condom that was p laced bet ween the standoff pad and sp onge p hantom . T he thick ness of gel in the cond om was m easured using u ltrasou nd imaging an d also tak ing p ictures with a r uler plac ed as ide as ref erence. T he thick ness of ultrasound transm iss ion gel was varie d at 4 levels: 0 (base) , 2 mm (level 1) , 7 m m (level 2) and 12 m m (level 3). A sinusoida l vibr ation signa l of 0.1 s dura tion was generat ed b y a f unction ge nerator (Mo del 3312 0A, Agilent, Santa Clara, C A). The vibration s ignals were used at fiv e f requenc ies of 100, 150, 200, 2 50 , and 3 00 Hz [13] . T he excitation s ignal at a fr equenc y was am plified b y an audio am plif ier (Model D 150A, Crown Aud io Inc ., Elk hart, IN) . T his s igna l then dr ove an e lectrom agnet ic shak er (Model: F G - 1 42, Labwork s Inc., C osta Mesa, CA 92 626) m ounted o n a stand. T he s hak er applied a 0. 1 s harm onic vibration o n the s urfac e of the ac oust ic stan doff pad using an indent er with 3 mm diam eter. 0.1 s is selected to exc lude m ost of the ref lections from the data col lection windo w whi le keepi ng the detec ted wave as a cont inu ous wav e. T he propagat ion of the v ibration wave in th e sponge was meas ured using a linear arra y tra nsducer ( L1 1 - 5v , Ph ilips H ealthcar e, Ando ver, MA) trans m itting at default 6.4 MH z center frequenc y mounte d on the ac oustic s tandof f pad. T he trans ducer was con nected to th e ultras ound system (Vantage 1, V erason ics Inc , Kir kland, W A) (Fig. 1) . T he m eas urements were repeated thre e tim es at eac h frequenc y and eac h gel thick ness . 6 Fig . 1. Exper im ental setup of sponge phantom , tr ansmis sion gel in a con dom , and ac oustic s tandof f pad. S tatis tica l an al ysis A n unpaired, t wo - tai led t - t est of the diff erences in s urfac e wave speed of the sponge phantom among dif fer ent levels of g el thic knes s was c onduct ed to com pare sam ple m eans. D iff erences in m ean values were cons idere d signif icant at p < 0.05. Numerical mo deling A FEM m odel was devel oped in ABAQUS (VERSIO N 6.14, 3D S Inc, W altham, MA). T he s ponge phantom , ac oustic standof f pad , and transmission gel in the condom were si m ulated as a 2D plan ar model of elastic m edium (Fig. 2) . L e ngth and h eight of the acoustic standof f pad were 9 and 1. 5 cm . Length and hei ght of the sponge phant om w ere 12 and 2 cm . U ltras ound trans m ission gel thick ness was predefin ed based on the m easur ement s fr om exper iments . T he de nsit ies of the acous tic stand off pad and transm iss ion gel were ass um ed to be 1000 k g m -3 . T he str uctural const ituent of the sponge p hantom is cellu lose , whic h has a dens ity of 1500 k g m -3 . T he s ponge was m odelled us ing a line ar por o - viscoelast ic model as suming the v oid ratio of the sponge is 0. 7 . The s tandoff pad was assum ed an incom pressible, linear elas tic m ater ial [14] . T o the best of our k no wledge, viscoelas tic pr operties of the spo nge ph antom have not be en qua ntifie d . T he values of the m ater ial param et ers of the sponge phantom were est im ated based on the ex perim ental m easurem ents in ter m s of surf ace wave speed of t he sponge p hantom at sponge gel pad shaker probe 7 differ ent vibrat ion f requen cies. G iven th e Voi gt m odel has widel y bee n used f or char acteri zing the dynam ic behavior of sof t biological t issues , t he Voi gt m odel  (  ) =   +     is used in t his s tud y . T he surf ace wave sp eed us ing t his m odel was c alcul ated a s:   =   .            󰇧            󰇨 ( 1) where   ,  ,   and   are surf ace wave spee d, angu lar fr equenc y , shea r ela stic it y , and s hear visc osity , respectiv ely . G ive n surf ace wav e spee ds of the sponge at five vibration f requencies ,   and   were identifie d via l east - s quare r egression and used to ca lculate stor age m odul us   , los s m odulus   , and long ter m m odulus   ,   =   (2)   =   ( 3)   =   ( 4) Thes e quantities can t hen be im plem ented in the A BAQU S in the fr equency dom ain with f r equenc y ranging f rom 100 to 3 00 H z at a n interv al of 50 Hz (Table 1) . Table 1. Materia l param eter s of s ponge phant om , acoustic standof f pad and ultr asound tr ans mis sion gel. Mat erials Sponge Standoff pad Transm ission gel E [kPa] 36.7 v 0.499   [kPa] 6.83 1.3   [Pa·S] 24 24 8 Fig . 2. Geom etric al conf iguration of s ponge p hantom m odel. T he m odel was ex cited usi ng a segm ent sour ce on t he top left surf ace of s tandof f pad and the displacem ent was app lied in the vert ical direct ion (Fig. 2) . H arm onic ex citat ions were per form ed at 100, 150 , 20 0, 250 , and 3 00 H z with duration of 0.1 s . T he c entral segm ent of the aco ustic s tandof f top s urfac e pad and bot tom surface of sponge were fixed in t he horizo ntal an d vertical directio ns. T he bottom boundar y of the sp onge p hantom wa s a ttach ed to an inf inite reg ion to m inim ize the wave reflect ions [ 15] . T he m esh of sponge , s tandoff pad , and u ltrasou nd transm is sion gel i n the condom we re construc ted us ing lin ear quadr ilater al plan e stres s e lements (t ype CPS4R) wit h size 1 mm x 1 mm , enhanced with hour glass control and reduc ed integr ati on, to m inimize shear lock ing an d hourglass eff ects. T he infinit e reg ion was m es hed b y infinite elem ents ( type CIN PE4) with size 1 mm x 1 mm . T he dynam ic responses of the s ponge p hantom m odel to the ex cit ations were solved by the ABAQU S implic it dy namic solver with a utom atic step s ize control. M esh con verg ence tests were perf orm ed so that furt her ref inin g the m esh did not c hang e the solu tion sig nifican tl y. pad gel sponge vibration fixed 9 R esults D etection of s ponge m otion is gu ided b y ultras ound im aging. S po nge m otion at a g iven loca tion can be an al yzed b y cross - correlatio n anal ysis of the ultr asound trac king be am . I n this stu dy, e ight locations on th e surf ace of the sponge phan tom over a lengt h of approxim atel y 8 mm were select ed to meas ure sponge m otion (Fig 3 a) . A hi gh fram e r ate of 2000 f rame/s is used to detec t sponge m otion i n response to th e vibrati on ex citation at 10 0, 150 , 200, 250 , a nd 3 00 H z. T he s urfac e wave speed is shown with 95% conf idence interv al, m ean ± s tand ard err or ( Fig. 3b). Fi g. 6a sh ows t he r elations hip bet ween surf ace wave s peed of sp onge p hantom and gel thi cknes s. S urf ace wav e sp eed ( SW S ) of the s ponge phantom increas ed with fr equenc y at eac h leve l of g el thic knes s, f rom 3. 28 ± 0.08 m /s at 100 Hz to 7.43 ± 0.19 m / s at 300 Hz. A t the sam e frequenc y, no statistica lly s ignific ant dif ferenc e in surf ace wave spee d of the sponge phantom was seen at dif ferent le vel s of gel thick ness . Fig. 3. (a) Re present ative B - m ode im age of sponge p hantom m odel. E ight locat ions on the surf ace of the spong e phant om were s elected to m easure t he wav e spe ed in the sponge b y using t he ultr asound tr ack ing beam m ethod. B lue dots indicat e the po ints se lected f or meas urem ent. (b) Repres entativ e phase del a y - distan ce relations hips of the spo nge at frequ enc y of 100 H z. T he wave phas e ch ange with posit ion, in respons e to a 0.1 vibrati on was us ed to m easure the wave speed. Standoff pad Ultrasou nd gel Sponge (a (a) (b) 10 FEM anal ysis of the sponge model was us ed to num eric ally i nvestig ate the ef fec t of transm iss ion gel thickness on the surf ace wav e speed of the spo nge phan tom . R es ults of the f inite elem ent sim ulations are sum m arized in figure 4 - 6. A s the boundar y of the sponge was assi gned inf inite e lement s, there is no wave ref lection in the boun dary (F ig. 4) . T he tem poral - s patial d isplacem ent f ield of a cent ral segm ent of sponge was ex tracted to m inim ize the inf luence of boundar y eff ects . The phase veloci ty can be m eas ured using 2D Fo urier trans form using th e fast F ourier trans form (FFT) on the spatiotem poral m otion data. T he resulting F ourier dis tribut ion, or k - spac e, has one tem poral f requenc y ( F ) axis and one spat ial f requenc y (K ) axis. 2D - FFT of the dis placem ent vers us t ime data was per form ed usin g   (  ,  ) =     (  ,  )   (     )       ( 5) where   (  ,  ) is the m otion of the sponge to the excitat ion as a func tion of distanc e from the exc itation ( x) and tim e (t). H ere, K is the wave n um ber and F is th e tem poral f requenc y of the wave . The c oordina tes of the k - space are t he wave n um ber (K) and the f requenc y (F ) (Fig. 5) [16, 17] . F or t he har m onic wave cas e, a peak will oc cur at the ex citati on f requenc y, and the c oordinates wher e th e pea k oc curs can be us ed to determ ine the ph ase ve locity us ing   =     (6) where c  is th e surf ace wave s peed, f  is the p eak t em poral f requenc y and k  is the peak spatial frequenc y. 11 Fig . 4. Conto ur of transla tional dis placem ent f ield of sponge ph antom model due t o harm onic vibr ation. Fig . 5. R epres entati ve k - s pace f rom 2D F FT transform ation of t he spon ge at 10 0 H z excitati on frequ enc y. The res ults f rom the num erical sim ulat ion sho wed th at th e surf ace w ave sp eed (SW S) of the sponge ph antom incr eased with exc itatio n freque nc y at diff erent l evel s of gel thick nes s (Fig. 6b). A t 1 00 and 150 H z, there was n o s ignifican t diff erence in t he sur fac e wave speed of the spon ge phan tom at differ ent level s of gel th ick ness . However, at 200, 250, and 300 H z, there was increas e in surf ace wa ve 12 speed b y 10% at u ltrasou nd transm is sion gel th ick ness of 2 mm , 15% at 7 m m and 35% at 12 mm relative to the surf ace wave s peed at b ase lev el of ultr asound gel th ick ness (Fig. 6b). Fig . 6. (a) S urf ace w ave s peed – gel th ick ness level s of sponge phantom at mult iple fr equenc ies (10 0, 150, 200, 250 , and 300 Hz) f rom 3 r epetitive m eas urem ents. E rr or bars r epresen t the sta ndard de viation of 3 rep etiti ve meas urem ents. (b) S urf ace wave spe ed – gel thick nes s levels of s ponge ph antom at multiple f requenc ies ( 100, 150, 200, 25 0 , and 300 H z) fr om num erical sim ulations . (a) (b) 13 D iscussio n s T he aim of this stud y was to develop a l ung s ponge phan tom m odel to investig ate the ef f ect of pleur al f luid o n surf ace wave prop agatio n in L U SW E . A s p onge ph antom , acoustic standof f pad , and ultrasound t ransm iss ion gel were simulated as lung parenc hyma, tho racic muscle , and ple ural flu id. T he level of transm iss ion gel th ick ness was adjus ted b y squee zing different amount of transm iss ion gel into a condom that was placed b etween th e ac oust ic s tandof f pad and s ponge phan tom . A t each leve l of gel thick ness, a shak er was used to ge nerat e a harm onic mec hanical vibr ation on th e acoustic standof f pad at one fr equenc y o f 100, 150 , 200, 2 50 an d 300 Hz. T he resu lting wa ve propaga tion on t he spong e surf ace was meas ured usi ng an ultras ound trans duc er. A FEM m odel was de veloped ac cord ing to the experim ental set up to sim ulate the wave propa gation on th e surf ace of the spon ge phant om . W e found influenc e of u ltrasou nd tran sm ission ge l thicknes s on the surface wave s peed of the spong e insignificant at fiv e diff erent ex citation fr equencies f rom experim ents . I n FEM, at 100 a nd 150 Hz of excitat ion frequenc y, no s ign ific ant influe nce of ultras ound trans m ission gel th ick ness on the s urfac e wave speed of the sponge was obser ved ye t at 200, 250 an d 300 H z, the surf ace wave speed o n the sponge i ncreas ed with thic knes s of ultrasoun d transm iss ion gel in F EM sim ulation. The dynamic respons e of the sponge phan tom obtained i n our stu dy has qu al itat ivel y sim ilar trends with that of lung p ar enchyma in vivo , as b oth e xhibit increasi ng shear wav e spe ed with exc itation frequenc y [1 8] . T he magnit ude of sur face wave spe ed of the sponge, which is an indicator of elasticit y of the s ponge , is hig her than that of health y subjec t s and sim ilar with that of pat ients with int erstiti al lung disease [7, 8, 19] . T he pre sented m odel is of partic ular releva nce to hum an m odel s a nd this techni que may be ada pted to asses s pati ents with pleur al ef fusion. L ung ultr asonogr aph y is m ore s ensitive, sp ecif ic , and repro duci ble f or diagnosi ng lu ng pathologi es an d can be co nsidere d an a lternati ve to bedsi de chec k radiograph y and thor acic c om puted tom ography [3] . I t can be us ed eff icientl y to eval uate the lung since m or e than 70% of the lun g can be imaged thr ough int ercosta l spaces [20] . L ung ultr ason ograp hy is exc ellent f or diagnosi ng pleura l diseas es, and it is espec iall y useful in th e em ergenc y and cr itical car e set tings f or t he det ection of pl eural ef fus ions or guidance of proc edures such as thoracent esis [21, 22] . The vo lume of pleural flui d has been estim ated 14 via ultras onograp hy for m echanicall y ventilate d patie nts or p atients with thoracen tesis [1 , 23] . C haracteri zation of elastic ity of superf icial lun g parenc hyma is crit icall y import ant for evalu ating the function of lung parenc h y m a [19, 2 4] . I n L U SW E , the surf ace wave on th e lung is safel y generated b y a local m echanica l vibration on the c hest. D iagnos tic ultr asound is only use d for detect ing surf ace wave propagat ion on the lung. H ence, it is a noninv asive a nd s afe techn ique f or lung te sting. W ith LUS WE , we plan to evaluate patien ts with pleura l eff usion via asse ss ing elastic proper ties of superf icia l lung parench yma pri or and af ter drainag e of pl eural f luid. F inite elem ent m odeling n um ericall y evaluated the ef fec ts of ultrasoun d trans mis sion gel on t he wave prop agatio n of the spong e surf ace. T he ultrasound tr ans m ission gel us ed to s im ulate p leu ral f luid that occup ies the pleur al c avity is a gel - lik e substanc e. G ive n its m ec hanical beha vior lies be tween Hookean s olid and N ewto nian f luid, the pleur al flui d was m odeled as a visc oelas tic m ateria l with high visco sit y . T he inf inite e lem ents ass igne d on the boun dar y of m usc le and lung wer e used t o get rid of the wave reflection , and t heref ore, im pr ove the ac curac y of wav e spe ed calc ulatio n. I t show ed that the wave speed incr ease d with exc itat ion frequ ency, which is in agreem ent with the obta ined ex perim ental meas urem ents. I n this study, th e s ponge was assum ed as a linear, poro - v iscoela stic m ater ial with a vo id ratio of 0.7 given i ts sim ilar st ructure with lung p arench yma [25 ] . A t higher frequ enc ies, an incr ease in t he surf ace wave spee d of th e sponge phant o m at different lev el s of gel thick ness could be due to a sim plified assum ption that t he s pong e phantom is a hom ogeneous mater ial with v oid unifor m ly distributed in the s ponge without a ir trapped within it . T he ex perim entall y m eas ured sh ear wave s peeds at f ive excitation fr equencies of t he sponge surf ace was use d to identif y its viscoel astic ity and the Voi gt model was used . D if ferent v iscoe lastic m odels m ay yield d ifferent r esults , and their pr edicti ons w ill be s tudied i n a separat e future w ork . W e hold that this inv estigat ion repr ese nts a m eaningful contribu tion to t he collect ive k nowledge of LU SW E , but ack nowledg e that there ar e lim itat io ns to the stud y that m erit fur ther consi derat ion . Intercost al muscle was simulat ed usin g an ac oustic s tandof f pad in this stu dy. S t ructur e and m echanic al propert ies of i n tercos tal muscle is m ore c omplicat ed than an acous tic sta ndof f pad. T he m easured surf ace wave s peed o n the lung sh ould have som e c ontribu tions f rom the int ercos tal mus cle. L u ng s lid ing 15 has been s imulate d us ing wet f oam as lung par ench yma and hand as thorac ic wall [10, 2 6] . Furt her impr ovement ma y use hand as t horac ic wall or dev elop an ex - v ivo anim al m odel to investigat e the ef fec t of pleura l fluid o n the wave propagat ion of lung par enc hyma in L U SW E . C onclusion I n sum m ar y , the prese nt m anusc ript dev elops a l ung spon ge phant om model to i ntegr ate experim ental m easurem ents and num eric al sim ulation to charac terize the ef fec t of pl eural f luid on t he dynam ic res ponse of t he s ponge phantom in t erm s of s urfac e wave sp eed. B oth the ex perim ents an d the FEM ana lyses s howed that u ltraso und transm iss ion ge l thick ness has an ins ignific ant ef fect on t he s urf ace wave spe ed of t he s ponge p hantom at ex citation f requenc ies of 100 and 150 H z . T he lung sponge p hantom m odel ma y be usef ul f or fur ther deve loping ex vivo anim al lung m odels an d in v ivo human lu ng s tudies. 16 A cknowledgements T his s tudy is support ed b y NIH R01HL12 5234 f rom the Nationa l Heart, Lu ng and Blood Inst itute. W e would lik e to thank Mrs . Jennifer Poston f or e dit ing this m anusc ript. Conflict of I nterest: The authors dec lare t hat the y ha ve no conf lict of inter est. 17 R eferences 1. Balik, M., et al., Ultrasound estim ation of volume of pleur al fluid in mechani cally ventilat ed patients. Intensi ve care medicine, 2006. 32 (2): p. 318. 2. Ahmed, S.H., et al., He modynamic and Pulm onary Changes af ter Drai nage of Significant P leural Effusions in Critically Ill, Mechanically Ventilated Surgica l Patients. Journal of T rauma and Acute Care Surgery, 2004. 57 ( 6): p. 1184 - 1188. 3. Licht enstein, D., e t al., Comparative diagnosti c performances of auscultation, chest radiography, and lung ultras onography in acute r espirator y distress syndr ome. Anesthesi ology: The J ournal o f the Ameri can S ociety of A nesthesiologis ts, 2004. 100 (1): p. 9 - 15. 4. Remérand, F. , et al., Multiplane ultrasound approach to quanti fy pleural effusion at the bedside. Intensive care medicine, 2010. 36 (4): p. 656 - 664. 5. Mayo, P.H . and P. Doel ken, Pleural ultrasonography. Clin Chest Med , 2006. 27 (2) : p. 2 15 - 27. 6. Soni, N.J., et al. , Ul tras ound in the Diagnosis & Management of Pleural Effusions . Journal of hospital medicine, 2015. 10 (12 ): p. 81 1 - 816. 7. Zhang, X., et al., Lung ultrasound surface wave elastogr aphy: a pilot clinical s tudy. IEEE transactions on ultrasonics, ferroelectrics, and frequency control, 2017. 64 (9) : p. 129 8 - 1304. 8. Kalra, S., et al. Lung ultras ound sur face wave e lastography - prel iminary measurement s in patients w ith inters titial lung diseas es . in Res pirology . 2017. Wil ey 1 11 River St, Hoboken 07 030 - 5774, NJ U SA. 9. Zhang, X., et al., An ultrasound surface w ave elastography techni que for noninvasive measureme nt of surface lung tissue. T he Journal of the Acoustical Society of America, 2 017. 141 (5): p. 3721 - 37 21. 18 10. Lee, K. - H., et al., Evaluation of a nove l simulat ion method of teaching B - lines: hand ultr asound with a wet foam dressing material. Clinical and experimental emergency medicine, 20 15. 2 (2) : p. 89. 11. Bl ü thgen, C., et al., Economical Sponge Phantom for Teaching, Underst anding, and Researchi ng A ‐ and B ‐ Li ne Re verberation Ar tifacts i n Lung Ultrasound. Journal of Ult rasound in Medicine , 2017. 36 (10): p. 2133 - 2142. 12. Cardenas - Garcia, J., P.H. M ay o, and E. Folch, Ultrasonographic evaluation of the pleura. Pleura, 2015. 2 : p. 2373 9975 156102 70. 13. Kubo, K., et al. , The quantitative e valuation of t he relationship betw een the forces applied to the palm and carpal t unnel pressure. Journal of Biomechanics. 14. Chen, L., et al., In Vivo Estimat ion of Perineal Body Pro perties Using Ultras ound Quasistatic Elastogr aphy in Nulliparous Wom en. Journal of biomec hanics, 2015. 48 (9) : p. 15 75 - 1579 . 15. Zhou, B. , A.J. Sit, and X. Zhang, Noninvasive measurement of wave speed of porcine cornea in ex vivo porcine eyes for various intraocular pressures. Ultrasonics, 2017. 81 (S upplement C): p. 86 - 92. 16. Bernal, M., et al., Material proper ty estimation for tubes and arter ies using ultrasound r adiation force and analys is of propagat ing modes. T he Journal of the Acoustical Society of America, 2011. 129 (3): p. 1344 - 13 54. 17. Kubo, K., et al ., Ultrasound elastography for carpal tunnel pr essure measurement: A cadaveric validation st udy. J Orthop Res, 2017. 18. Xiaoming Zhang, B.Z., Thomas Osborn, Brian Bartholmai, James Greenleaf, Sanjay Kalra, Assessment of interst itial lung disease using lung ultr asound surface wave elastography. Ultrasonics Symposium (IUS ), 2017 IEEE International, 2017: p. 1 - 4. 19 19. Osborn, T., et al. A Non - Invasive Ultr asound Surface Wave Elastogr aphy Technique for Assessing Interst itial Lung Diseas e . in ARTHRITIS & RHEU MATOLOGY . 2 017. WILEY 1 11 RIVER S T, HOBOKEN 07030 - 5774 , NJ U SA. 20. Mayo, P.H. and P. Doe lke n, Pleural ultraso nography. Clinics in chest medicine, 2006. 27 (2 ): p . 215 - 227. 21. Mathis, G., At las of chest sonography . 2003: Springer Science & Business Media. 22. Hakimisefat, B. and P.H. M ay o, Lung ultrasonogr aphy. The Open Critical Care Medicine Journal, 2010. 3 (2): p. 21 - 25 . 23. Vignon, P., et al., Quantitative assessme nt of pleural effusion in crit ically ill patients by means o f ultrasonogr aphy. Critical car e medicine, 2 0 05. 33 (8): p. 1757 - 1763. 24. Xiaoming Zhang, B.Z., Sanjay Kalra, Brian Bartholmai, James Greenleaf, T homas Osborn, Quantitativ e assessment of scle roderma usi ng ultrasound sur face wave elast ography. Ultrasonics Symposium (IUS ), 2017 IEEE Inter national, 201 7: p. 1 - 3. 25. Luo, H., I. Goldst ein, and D. Udelson, A t hr ee ‐ dimensiona l theoretical model of the relations hip between caver nosal expandability and percent cavernos al smoot h muscle. The j o urnal of sexual med ici ne, 2007 . 4 (3) : p. 644 - 655. 26. Shokoo hi, H. and K . Boniface, Hand Ul trasound: A High ‐ fidelity Simulation of Lung Sl iding. Academic Emergen cy Medicin e, 2012. 19 (9). 20