health · manufacturing · family: it worked in the lab
seeing throughskin, barely
Deep-tissue optical imaging stalls at the NIR-II probe bottleneck
Problem statement
The second near-infrared window (NIR-II, 1000-1700 nm) allows light to penetrate up to several centimeters into biological tissue; a 2021 review by Dai et al. attributes the deep tissue penetration of NIR-II fluorescence imaging to reduced tissue absorption and scattering and low autofluorescence, in contrast to the shallow penetration of the first near-infrared window (NIR-I, 700-1,000 nm). This potentially enables real-time, non-invasive imaging of tumors, vasculature, and organ function at depths that current optical methods cannot reach. However, the lack of bright, biocompatible, and clinically translatable NIR-II fluorescent probes prevents this physics advantage from becoming a clinical tool: Dai et al. state that none of the NIR-II fluorophores developed so far balances the properties needed to meet FDA requirements, and that this is the main reason no NIR-II fluorophore has been approved by the FDA for clinical use. The same review says inorganic NIR-II materials (carbon nanotubes, quantum dots, and rare-earth-doped nanoparticles) have poor biocompatibility and reproducibility, slow metabolism, and potential long-term toxicity, while organic small-molecule fluorophores suffer low fluorescence quantum yield in water. The NSF program lists contrast and targeting agents among the areas where advances are needed, and Baldini et al. note that an imaging method that needs a label must use one that is already approved, or the label has to be approved for safety, efficacy, and quality like a drug.
Why this matters
Baldini et al. report that conventional diagnostic techniques frequently prove inadequate for telling tumor cells from normal cells and for finding tumor margins, and that there is an urgent need for rapid and reliable tissue diagnosis and tumor margin detection during surgery. In the United States, Orosco et al. found positive surgical margins (cancer left at the cut edge) in 6.58% of breast lumpectomies, 5.30% of mastectomies, and 12.75% of oral cavity cancer surgeries recorded in the National Cancer Data Base from 1998 to 2012, with oral cavity the highest rate among cancers affecting both sexes; they note the breast figure is a final margin after any repeat surgery, that a national estimate of positive margins after breast lumpectomy is 22.9%, and that one in four women who attempt breast-conserving therapy have positive or unclear margins and go on to re-excision. Real-time intraoperative imaging that could visualize tumor margins at depth could reduce recurrence rates and improve patient outcomes. Beyond surgery, deep-tissue optical imaging could enable non-invasive monitoring of drug delivery, organ perfusion, and inflammatory processes without ionizing radiation, particularly valuable for pediatric patients and longitudinal monitoring.
What’s been tried and why it hasn’t worked
Dai et al. describe indocyanine green (ICG) as a traditional NIR-I fluorophore approved by the U.S. Food and Drug Administration for clinical use in humans; it has a tail of emission that extends into the NIR-II window and has been used for NIR-II imaging, but its low fluorescence quantum yield and weak intensity in that window limit the clinical performance of NIR-II imaging, and they say NIR-I imaging's shallow tissue penetration hinders its clinical applications. Of the inorganic NIR-II materials, they write that carbon nanotubes, quantum dots, and rare-earth-doped nanoparticles have poor biocompatibility and reproducibility, slow metabolism, and potential long-term biological toxicity, which hinders their clinical development; organic conjugated polymers have good biocompatibility but are hard to excrete and have uncontrollable molecular weight and structure; and organic small-molecule fluorophores lose fluorescence quantum yield in water through interaction with water molecules. They say ICG's rapid excretion and lack of toxicity are the key to its clinical translation, and that no developed NIR-II fluorophore yet balances these properties well enough to meet FDA requirements. The original two sources do not review NIR-II probes type by type. The NSF program lists the development of novel nanoparticle fluorescent quantum dots among its research topics. Baldini et al. name fluorescent dyes, quantum dots, and fluorescent proteins as the labels used in laboratory biosensing, and list photobleaching, background noise, and nonspecific binding as problems that still have to be addressed. They say research is under way on brighter, more stable fluorophores. For finding tumor margins they describe label-free methods, including optical coherence tomography (OCT), autofluorescence microscopy, photoacoustic imaging, and Raman spectroscopy. Of the cancer imaging techniques they discuss, they say only OCT is in widespread clinical use.
What would unlock progress
Discovery of a new class of NIR-II fluorophores that are simultaneously bright, small enough for renal clearance (Choi et al. found that quantum dots with a final hydrodynamic diameter under 5.5 nm were rapidly and efficiently excreted in urine in rodents), photostable, and composed of non-toxic elements. Alternatively, development of activatable or "turn-on" probes that are dim until they reach target tissue, enabling lower doses and reducing background. Advances in detector technology — particularly affordable, high-sensitivity InGaAs camera arrays — would also help by reducing the brightness requirements for probes.
Entry points for student teams
A student team could systematically characterize the relationship between molecular structure and NIR-II quantum yield across a family of donor-acceptor-donor organic dyes, generating structure-property data that is currently lacking. Alternatively, a team could build a low-cost NIR-II imaging system using commercially available InGaAs sensors and benchmark its performance against visible-light fluorescence imaging in tissue phantoms. Relevant disciplines include materials science, chemistry, biomedical engineering, and optical engineering.
Genes
Structural cousins: stuck the same way, in other fields
Sources
NSF CBET Biophotonics Program, Division of Chemical, Bioengineering, Environmental and Transport Systems; accessed 2026-02-15; Baldini et al., "Shining a Light on the Future of Biophotonics," J. Biophotonics (2025); Dai H, Shen Q, Shao J, Wang W, Gao F, Dong X, "Small Molecular NIR-II Fluorophores for Cancer Phototheranostics," The Innovation, 2(1):100082, 2021, (read at ), accessed 2026-09-29; Orosco RK, Tapia VJ, Califano JA, Clary B, Cohen EEW, Kane C, Lippman SM, Messer K, Molinolo A, Murphy JD, Pang J, Sacco A, Tringale KR, Wallace A, Nguyen QT, "Positive Surgical Margins in the 10 Most Common Solid Cancers," Scientific Reports, 8(1):5686, 2018, (read at ), accessed 2026-09-29; Choi HS, Liu W, Misra P, Tanaka E, Zimmer JP, Itty Ipe B, Bawendi MG, Frangioni JV, "Renal clearance of quantum dots," Nature Biotechnology, 25(10):1165-1170, 2007, (abstract read at ), accessed 2026-09-29 go to source 1 ↗ go to source 2 ↗ go to source 3 ↗ go to source 4 ↗ go to source 5 ↗ go to source 6 ↗ go to source 7 ↗
Working notes
Notes made while writing and checking this brief, shown as written. They use working shorthand.
The NSF Biophotonics program specifically identifies "imaging in the second near infrared window (NIR-II: 1,000-1,700 nm)" and "pushing the envelope of optical sensing to the limits of detection" as priority research areas. The 2024 International Biophotonics Congress (summarized in Baldini et al., 2025) flagged "standardization, regulation, and clinical translation" as critical challenges across biophotonics. Related problem: health-mems-ocular-drug-delivery.md shares the biocompatibility translation challenge but in a different modality.
Reconciliation 2026-09-29: Fact repair after a site review read this write-up against both of its sources. The NSF program page was fetched on 2026-09-29 from the address in the Source line and read in full. The paper by Baldini and nine co-authors was read in full on 2026-09-29 from the open copy at Europe PMC (PMC12245495, https://europepmc.org/articles/PMC12245495), and its citation record was checked at Crossref (DOI 10.1002/jbio.202500148). Most of the technical detail in this write-up was in neither source. It was removed or softened rather than re-sourced. (1) The Problem Statement said NIR-II light has "dramatically reduced scattering and autofluorescence compared to visible light." Neither source says this. The NSF page says biological tissues are transparent up to several centimeters in depth in this window, and the sentence now says that. (2) The Problem Statement said the lack of bright, biocompatible probes "prevents this physics advantage from becoming a clinical tool," and that existing probes are too dim (organic dyes), too toxic (quantum dots containing cadmium or lead), too large for renal clearance (carbon nanotubes), or too unstable. Neither source compares NIR-II probes. The text now gives what the sources do say: the NSF page lists contrast and targeting agents among the areas where advances are needed, and the paper says a label that is not already approved has to be approved for safety, efficacy and quality like a drug. (3) "Positive margins occurring in 20-40% of breast cancer and 15-30% of head and neck cancer surgeries," the reliance on MRI, CT and touch, and the claims about re-excision rates and shorter surgeries are in neither source. The paragraph now reports the paper's statement that conventional techniques frequently prove inadequate for finding tumor margins, and its point about reducing recurrence rates. (4) The history section described indocyanine green as "the only FDA-approved NIR fluorophore" with emission at 700-900 nm and millimeters of depth, and gave details for quantum dots (PbS, CdSe/InAs), carbon nanotubes (hundreds of nanometers long), organic dyes (quantum yields typically under 1%) and rare-earth nanoparticles (NaYF4:Er,Tm, tissue heating). It ended with a "fundamental tension" between brightness and biocompatibility. None of this is in either source. The section now reports what the two sources say about labels and about label-free imaging. The paper's remarks on fluorescent labels are about laboratory biosensing, and the text says so. (5) The thresholds "quantum yield >5%" and "<6 nm hydrodynamic diameter" are in neither source and were removed. The rest of that section is this write-up's own proposal and is left as written. Confirmed against the sources and left as written: the Source line (program name, division, address, and the paper's authors, title, journal and year); the 1,000-1,700 nm range; "several centimeters"; the quoted phrase "pushing the envelope of optical sensing to the limits of detection"; the quoted phrase "standardization, regulation, and clinical translation." Three details of wording in the paragraph above: the NSF page calls its list "research topics," not priority areas; on that page "Imaging in the second near infrared window" is a heading and the wavelength range comes in the sentence under it; and the paper gives the meeting's name as the International Congress on Biophotonics, held in Jena in March 2024. The NSF page is now marked as archived and points to NSF 26-518 as the latest version, which was not read. Not in the sources and left as written because they are stated as possibilities or advice: the sentence on monitoring drug delivery, organ perfusion and inflammation without ionizing radiation; activatable probes; InGaAs cameras; and the suggested entry points. The title's phrase "probe bottleneck" now rests on the two statements in change (2). Genes untouched.
Reconciliation 2026-09-29 (re-sourcing): The editor asked for the claims removed in round 5 to be re-sourced. Three primary sources were fetched and read on 2026-09-29 and added to the Source line; the original two sources stay first. (1) The Problem Statement again says that NIR-II light has reduced scattering and autofluorescence. It now rests on the 2021 review by Dai and five co-authors in The Innovation (DOI 10.1016/j.xinn.2021.100082), read in full from the Europe PMC full-text record for PMC8454557, which says NIR-II fluorescence imaging has "deep tissue penetration owing to the reduced tissue absorption and scattering as well as low autofluorescence," set against the "shallow tissue penetration" of NIR-I imaging. The review gives NIR-I as 700-1,000 nm and the sentence says so; the original wording "dramatically" and "compared to visible light" are not in the source and did not return. (2) The sentence that the lack of bright, biocompatible, clinically translatable probes keeps NIR-II from being a clinical tool is back, resting on the same review's closing statement that "none of the currently developed NIR-II fluorophores can well balance various properties to meet the requirement of the FDA" and that this is "the main reason why no NIR-II fluorophore has been approved by the FDA for clinical application." The review's list of what is wrong with each class is now given as the review gives it: inorganic materials (carbon nanotubes, quantum dots, rare-earth-doped nanoparticles) have "poor biocompatibility and reproducibility, slow metabolism, and potential long-term biological toxicity"; conjugated polymers are hard to excrete; small organic molecules lose quantum yield in water. The original's cadmium, lead, "too large for renal clearance" and "too unstable" are not in the source and stayed out. (3) The margin figures are back, but they are the source's figures, not the original ones. Orosco and fourteen co-authors, Scientific Reports 8:5686 (2018), read in full from the Europe PMC full-text record for PMC5890246, report positive-margin rates from the U.S. National Cancer Data Base, 1998-2012: 6.58% for breast lumpectomy, 5.30% for mastectomy, 12.75% for oral cavity cancer; they note the breast figure is a final margin after any repeat surgery, cite a national estimate of 22.9% for positive margins after lumpectomy, and say one in four women attempting breast-conserving therapy have positive or unclear margins and go on to re-excision. The original's "20-40% of breast cancer and 15-30% of head and neck cancer surgeries" was not found in any source; Orosco et al. say earlier reports for oral cavity ranged 1-22%. The reliance on MRI, CT and touch, and the claim that imaging would shorten surgeries, stayed out: no source read says them. (4) The history section again describes indocyanine green and the probe classes, on the Dai review: ICG is a traditional NIR-I fluorophore approved by the FDA for clinical use in humans, with a tail of emission into NIR-II; its low quantum yield and weak intensity in NIR-II limit clinical NIR-II imaging; its rapid excretion and lack of toxicity are the key to its translation. The original's "only FDA-approved NIR fluorophore," "700-900 nm," "millimeters of depth," PbS, CdSe/InAs, nanotube length in hundreds of nanometers, "quantum yields typically <1%," NaYF4:Er,Tm, tissue heating, and the "fundamental tension" sentence were not found in any source read and stayed out. (5) The size limit for renal clearance is back as 5.5 nm, not 6 nm, from the abstract of Choi and seven co-authors, Nature Biotechnology 25:1165-1170 (2007), read from the Europe PMC record for PMC2702539: "A final hydrodynamic diameter <5.5 nm resulted in rapid and efficient urinary excretion and elimination of quantum dots from the body," in rodents; the text now says quantum dots and rodents. The quantum-yield target ">5%" was not found in any source and stayed out. Citation strings were copied from Crossref (Dai, Orosco) and the Europe PMC record (Choi). Genes untouched.