Why is On­col­o­gy Drug De­vel­op­ment Re­search Late to the Dig­i­tal Bio­mark­ers Game?

Dur­ing the re­cent An­nu­al AS­CO Meet­ing, thou­sands of can­cer re­searchers and clin­i­cians from across the globe joined to­geth­er vir­tu­al­ly to present and dis­cuss the lat­est find­ings and break­throughs in can­cer re­search and care. There were more than 5000+ sci­en­tif­ic ab­stracts pre­sent­ed dur­ing this event, yet on­ly a hand­ful in­volved the use of mo­tion-track­ing wear­ables to col­lect dig­i­tal mea­sures re­lat­ing to ac­tiv­i­ty, sleep, mo­bil­i­ty, func­tion­al sta­tus, and/or qual­i­ty of life. Al­though these re­sults were a bit dis­ap­point­ing, they should come as no sur­prise to those of us in the wear­able tech­nol­o­gy field.

De­spite tremen­dous ad­vances in pre­ci­sion med­i­cine, im­munother­a­py, and tar­get­ed treat­ments over the last sev­er­al decades, on­col­o­gy drug de­vel­op­ment re­search has lagged be­hind oth­er ther­a­peu­tic ar­eas in the adop­tion of wear­able dig­i­tal bio­mark­er tech­nolo­gies, such as actig­ra­phy. In fact, of the 207 (and count­ing) dig­i­tal end­points list­ed in the Dig­i­tal Med­i­cine So­ci­ety’s (DiMe) crowd-sourced li­brary of dig­i­tal end­points, none are be­ing used in an on­col­o­gy study. While the use of these tech­nolo­gies has steadi­ly in­creased in CNS, car­dio­vas­cu­lar, im­munol­o­gy, and meta­bol­ic dis­ease stud­ies, the pace of dig­i­tal trans­for­ma­tion with­in on­col­o­gy re­search seems to have stag­nat­ed in com­par­i­son.

So what is it about on­col­o­gy stud­ies that sets them apart from oth­er drug tri­als? Af­ter pos­ing this ques­tion to sev­er­al ex­perts in the field and re­view­ing a pletho­ra of sci­en­tif­ic lit­er­a­ture and in­dus­try com­men­tary on this top­ic, I was able to cat­e­go­rize the bar­ri­ers that emerged in­to two broad ar­eas – study com­plex­i­ty and pa­tient pop­u­la­tion.

High­ly Com­plex Stud­ies

On­col­o­gy clin­i­cal tri­als are ex­ceed­ing­ly com­plex com­pared to those in oth­er ther­a­peu­tic ar­eas. Al­though the chal­lenges faced may not be unique to on­col­o­gy, they do seem to be mag­ni­fied with­in these stud­ies. Ac­cord­ing to a re­cent re­port from the Tufts Cen­ter for Drug De­vel­op­ment re­search, on­col­o­gy drug tri­als typ­i­cal­ly the fol­low­ing is­sues2:

  • Greater dif­fi­cul­ties with pa­tient re­cruit­ment and re­ten­tion
  • Length­i­er screen­ing and treat­ment du­ra­tions
  • More com­pli­cat­ed study de­signs
  • More fre­quent pro­to­col de­vi­a­tions and amend­ments
  • More in­ves­tiga­tive sites across a wider glob­al foot­print

This begs the ques­tion, is the lag­ging adop­tion of wear­ables and oth­er dig­i­tal mea­sure­ment tech­nolo­gies with­in on­col­o­gy re­search root­ed in a hes­i­tan­cy on the part of spon­sors to add more lay­ers of com­plex­i­ty to an al­ready com­plex study? I would ar­gue that not on­ly is it pos­si­ble to col­lect dig­i­tal mea­sures like actig­ra­phy da­ta with­out adding sig­nif­i­cant chal­lenges, but in fact, these tech­nolo­gies can ac­tu­al­ly help mit­i­gate some of these is­sues in the first place.

Let’s take a look at pa­tient re­cruit­ment and re­ten­tion, which is ar­guably the biggest chal­lenge spon­sors face across all drug stud­ies. Ac­cord­ing to the DiMe So­ci­ety’s Phar­ma Ex­ec Mi­cro-Play­book, when used as a screen­ing tool, dig­i­tal clin­i­cal mea­sures can help speed re­cruit­ment by al­low­ing ac­cess to a larg­er pool of po­ten­tial par­tic­i­pants and re­duc­ing de­pen­dence on clin­i­cal sites.3 High­ly pre­cise dig­i­tal mea­sures al­so re­duce the like­li­hood of rater bias, al­low­ing for more tar­get­ed en­roll­ment and an in­creased prob­a­bil­i­ty of suc­cess.3

These oth­er chal­lenges – com­plex study de­signs, fre­quent pro­to­col de­vi­a­tions, and a larg­er glob­al foot­print – should not be a bar­ri­er to tech­nol­o­gy adop­tion. Rather they should be used to de­ter­mine whether a wear­able tech­nol­o­gy part­ner’s prod­ucts and ser­vice of­fer­ings are aligned with the study re­quire­ments. For ex­am­ple, Acti­Graph’s flex­i­ble, fit-for-pur­pose tech­nol­o­gy ecosys­tem was specif­i­cal­ly de­vel­oped to ac­com­mo­date a va­ri­ety of da­ta col­lec­tion and op­er­a­tional work­flows for com­pli­cat­ed study de­signs and lengthy da­ta col­lec­tion pe­ri­ods. Ded­i­cat­ed project man­agers han­dle down­stream im­pacts of pro­to­col changes to work­flows, pa­tient-fac­ing ma­te­ri­als, and site train­ing re­sources, while an in-house lo­gis­tic team with glob­al ex­per­tise man­ages site ship­ments and study in­ven­to­ry.

Unique Pa­tient Pop­u­la­tion

Can­cer pa­tient pop­u­la­tions can al­so present unique chal­lenges with­in the con­text of clin­i­cal re­search. Un­like most oth­er ther­a­peu­tic ar­eas, there of­ten ex­ists both a symp­tom bur­den and a treat­ment bur­den in can­cer pa­tients. The symp­tom bur­den could be vir­tu­al­ly nonex­is­tent, while the treat­ment bur­den – chemother­a­py side ef­fects, for ex­am­ple – takes a heavy toll. This can present chal­lenges in the way clin­i­cal teams as­sess treat­ment ef­fi­ca­cy and qual­i­ty of life in these pa­tients.

In most oth­er ther­a­peu­tic ar­eas, in­creased ac­tiv­i­ty lev­els dur­ing an in­ter­ven­tion can be an in­di­ca­tor of ther­a­peu­tic ef­fi­ca­cy. If a treat­ment is work­ing and the pa­tient is feel­ing bet­ter, they tend to move more, right? But in a can­cer study, ac­tiv­i­ty lev­els are like­ly to ac­tu­al­ly de­crease dur­ing treat­ment due to the side ef­fects of these pow­er­ful drugs. In this case, com­par­ing the ac­tiv­i­ty lev­els of pa­tients re­ceiv­ing the in­ves­ti­ga­tion­al prod­uct against those re­ceiv­ing the stan­dard of care treat­ment may pro­vide more mean­ing­ful in­sights in­to their qual­i­ty of life. Mea­sur­able im­prove­ments to qual­i­ty of life, even when sur­viv­abil­i­ty re­mains un­changed, may be enough to help bring an in­ves­ti­ga­tion­al drug to mar­ket.

Study teams may al­so be re­luc­tant to “bur­den” these pa­tients, who are of­ten old­er, very sick, and may be fac­ing a ter­mi­nal di­ag­no­sis, with ad­di­tion­al tech­nol­o­gy com­po­nents. Re­search sug­gests that these con­cerns may be mis­guid­ed. A re­cent ac­celerom­e­ter ac­cept­abil­i­ty and fea­si­bil­i­ty study of breast can­cer pa­tients un­der­go­ing treat­ment found that 100% of the par­tic­i­pants were con­fi­dent in their abil­i­ty to use the study’s tech­nol­o­gy and 78% rat­ed their study ex­pe­ri­ence as pos­i­tive.4 As in the case of study com­plex­i­ty, con­cerns about tech­nol­o­gy ac­cep­tance and bur­den should not be a de­ter­rent, but rather a guide to help spon­sors se­lect the most ap­pro­pri­ate wear­able de­vice for their pop­u­la­tion.

The col­lec­tion of dig­i­tal ac­tiv­i­ty mea­sures dur­ing on­col­o­gy drug tri­als can great­ly en­hance a spon­sor’s un­der­stand­ing of dis­ease pro­gres­sion, treat­ment im­pacts, and qual­i­ty of life. Yes, this ther­a­peu­tic area presents some unique chal­lenges, but I re­main very op­ti­mistic that the use of wear­able tech­nolo­gies with­in on­col­o­gy drug re­search will catch up with the rest of the in­dus­try.

Are you in­ter­est­ed in us­ing Acti­Graph for an up­com­ing clin­i­cal tri­al? Please con­tact us to dis­cuss your spe­cif­ic study needs with a mem­ber of our sci­en­tif­ic team.

1. https://www.dime­so­ci­ety.org/com­mu­ni­ca­tion-ed­u­ca­tion/li­brary-of-dig­i­tal-end­points/
2. https://www.globe­newswire.com/news-re­lease/2021/05/04/2222347/0/en/Pro­to­col-Com­plex­i­ty-and-Pa­tient-En­roll­ment-In­ten­si­fy-Chal­lenges-in-On­col­o­gy-Clin­i­cal-Tri­als-Ac­cord­ing-to-Tufts-Cen­ter-for-the-Study-of-Drug-De­vel­op­ment.htm
3. https://play­book.dime­so­ci­ety.org/play­books/mi­cro-play­books/
4. Solk P, Gavin K, Fan­ning J, Welch W, Lloyd G, Cot­trell A, Nielsen A, San­ta Maria CA, Gradishar W, Khan SA, Kulka­rni S, Sid­dique J, Phillips SM. Fea­si­bil­i­ty and ac­cept­abil­i­ty of in­ten­sive lon­gi­tu­di­nal da­ta col­lec­tion of ac­tiv­i­ty and pa­tient-re­port­ed out­comes dur­ing chemother­a­py for breast can­cer. Qual Life Res. 2019 Dec;28(12):3333-3346. doi: 10.1007/s11136-019-02278-7. Epub 2019 Sep 6. PMID: 31493269.