Theranostics in oncology: beyond finding the target

A woman with a small bowel neuroendocrine tumor and liver metastases undergoes a PET scan with a somatostatin analog. The lesions take up a lot, more than the normal liver. Someone in the meeting sums it up: "the target is there, she is a candidate". The sentence is right and, at the same time, tells only the beginning of the story.

Theranostics is the joining of diagnosis and therapy with the same molecule, or with sister molecules. A version labeled with an imaging emitter shows where the target is, and a version labeled with a therapeutic emitter delivers radiation to the same place. The image that selects the patient is only the first of several decisions, and each one of them requires clinical reasoning.

This text treats theranostics as a complete strategy. It goes through the four stages of the cycle: selecting by imaging, treating with the radioligand, measuring response by imaging and following toxicity. The examples come from the two most mature platforms, the somatostatin receptor (SSTR) one in neuroendocrine tumors and the prostate-specific membrane antigen (PSMA) one in prostate cancer.

Key points

• The target on imaging is necessary, but selection also looks at the disease that the target image may miss (Hindié, 2017).
• PSMA PET is the selection standard for lutetium-PSMA, with FDG, post-treatment SPECT and MRI as adjuncts (Ranjan, 2026).
• Response measured by PSMA PET, added to PSA, predicts survival better than PSA alone (Gafita, 2022).
• In neuroendocrine tumors, response should be assessed with the same modality as the baseline (Bodei, 2013).
• Kidney and bone marrow are the critical organs, and blood test follow-up continues after the last cycle (Bodei, 2013).
• Myeloid neoplasms can appear years after radioligand therapy (Strosberg, 2021).

Summary

• Why theranostics is a cycle, not a scan
• First stage: selecting by imaging
• Second stage: treating with the radioligand
• Third stage: measuring response by imaging
• Fourth stage: following toxicity
• Building the cycle step by step
• How NEXO helps train this reasoning
• Frequently asked questions
• Milestone inventory
• Take-home messages

Why theranostics is a cycle, not a scan

The idea is old. The use of radioactive iodine to treat thyroid disease, started by Saul Hertz, is considered the first theranostics, because the same substance that shows the gland also treats it (Oldan, 2025). Metaiodobenzylguanidine (MIBG), which is taken up in pheochromocytoma, paraganglioma and neuroblastoma, is described as the first agent developed on purpose to see and then treat (Oldan, 2025).

What has changed in recent years is the scale. Somatostatin analogs and PSMA ligands took theranostics to common diseases, to phase 3 trials and to the routine of oncology services. With the scale came a risk in reasoning: treating a positive scan as if it were the whole indication.

The joint guideline for lutetium-PSMA asks that the indication be discussed in an interdisciplinary tumor board and treats long-term follow-up as mandatory (Kratochwil, 2023). The guidance for peptide therapy in neuroendocrine tumors makes the same request for a multidisciplinary decision (Bodei, 2013). Both describe a cycle, not an isolated procedure.

Thinking in cycles changes the questions. At selection, the question is no longer "is there a target?" and becomes "does all the disease that matters have a target?". At treatment, the team asks what the person has already received and what the marrow and the kidney can withstand. At response, it asks which scan, at what time and with what criterion. At follow-up, it asks how long to monitor and what to look for. The decision of when to call nuclear medicine is part of this conversation from the start.

First stage: selecting by imaging

The target image is the gateway. In NETTER-1, which compared lutetium-DOTATATE with long-acting octreotide in midgut neuroendocrine tumors, somatostatin receptors had to be present in all target lesions, with confirmation by independent central review (Strosberg, 2017). Uptake was graded by the lesion with the highest uptake on somatostatin analog scintigraphy (Strosberg, 2017).

In prostate cancer, PSMA PET is the established scan for selecting candidates for lutetium-PSMA (Ranjan, 2026). The uptake criteria used in VISION and the marrow and kidney limits for indicating treatment are in the text on who is a candidate by PSMA PET, and will not be repeated here.

The disease that the target image can hide

The weak point of selection is in lesions that have lost the target. A tumor that dedifferentiates may stop expressing somatostatin receptors or PSMA and gain glucose metabolism. These lesions appear on FDG PET and disappear on the target image. If they are forgotten, treatment reaches the lesions that take up the tracer and lets precisely the most aggressive ones grow.

In neuroendocrine tumors, the NETPET score organizes this dual reading. It ranges from P1, when all lesions take up somatostatin analog and none take up FDG, to P5, when there are lesions that take up FDG and do not take up the analog (Hindié, 2017). In a study of 62 patients discussed in the editorial, the score was associated with survival (Hindié, 2017).

In prostate cancer, discordance has the same meaning. A retrospective comparison showed that men who met the stricter criteria of the TheraP study, which required the absence of FDG-positive lesions without PSMA, had more PSA response and longer progression-free survival, and that FDG-positive and PSMA-negative disease had a worse course (Ranjan, 2026). The joint guideline treats FDG PET as optional, useful especially for liver lesions (Kratochwil, 2023).

When reading the selection, some questions help decide: whether all lesions seen on CT or MRI take up the target tracer; whether there are large, liver or rapidly growing lesions that call for FDG; whether the histological grade and the speed of progression fit a well-differentiated disease; whether the marrow has enough reserve after previous treatments; whether kidney function allows a treatment that passes through the kidney; and whether the person can follow the radiation protection rules at home.

When PET is not available

Access to PET varies widely between services. The review on multimodal imaging describes SPECT with technetium-99m-labeled PSMA as a more accessible and lower-cost alternative, with comparable diagnostic performance (Ranjan, 2026). The same review points to lower detection in the prostatic bed with this technique, which weighs on the reading of local disease (Ranjan, 2026).

Target expression also changes with treatment. PSMA expression appears to increase in the first months of androgen blockade (Zhang, 2021). That is why the timing of the selection image relative to other treatments goes into the report and the team discussion.

Second step: treat with the radioligand

The radioligand binds to the same target as the image and carries a beta emitter, such as lutetium-177, to the cells that express it. In VISION, lutetium-PSMA-617 was described as a therapy that delivers beta radiation to cells expressing PSMA and to the surrounding microenvironment (Sartor, 2021). The radiation travels a few millimeters through tissue, which concentrates the effect in the lesions and explains why selection by imaging weighs so heavily.

The results on the two platforms show the size of the effect. In NETTER-1, 229 patients were randomized, and 116 received lutetium-DOTATATE (Strosberg, 2017). Estimated progression-free survival at month 20 was 65.2 in 100 with lutetium, versus 10.8 in 100 in the control group (Strosberg, 2017). The risk of progression or death was 79 in 100 lower (Strosberg, 2017). The response rate was 18 in 100, versus 3 in 100 (Strosberg, 2017).

In prostate cancer, VISION randomized 831 of the 1179 men screened (Sartor, 2021). What the study showed and what came afterward, with PSMAfore and PSMAddition, is in the text on what changed after VISION.

What accompanies the infusion

The infusion itself is the shortest part of the cycle, but it carries precautions that come from the biology of the target. In peptide therapy, the kidneys take up the somatostatin analog through a proximal tubule transport system, and amino acid infusion is used to reduce this uptake (Bodei, 2013). These amino acids explain much of the nausea and vomiting on the day of treatment: in NETTER-1, nausea appeared in 59 out of every 100 patients and vomiting in 47 out of every 100, largely attributed to the amino acid infusion (Strosberg, 2017).

The somatostatin analog also protects against carcinoid crisis in people with functional syndrome (Bodei, 2013). Flushing, diarrhea, bronchospasm and blood pressure instability on the day of infusion require immediate attention.

Individual treatment planning based on organ dosimetry is still research. For lutetium-PSMA, the joint guideline says that individual dosimetry is optional in standard use and that the clinical gain of dosimetry-guided regimens still needs confirmation (Kratochwil, 2023). An ongoing phase 3 randomized trial compared the standard regimen with strategies adjusted at each cycle by kidney and marrow dosimetry, and the interim analysis showed that these strategies are feasible and safe (Kolodziej, 2025). The question of survival gain remains open.

Third step: measuring response by imaging

Theranostics offers an advantage that other systemic therapies rarely have: the treatment itself can be photographed. The joint guideline describes scintigraphy or SPECT performed 1 to 2 days after lutetium-PSMA infusion, which show where the radioligand bound (Kratochwil, 2023). In peptide therapy, post-treatment images report the intensity and location of uptake and help assess response to previous cycles (Bodei, 2013).

PSA, flare and the right moment

PSA is the most used marker in prostate cancer, but it can mislead at the start. The joint guideline calls for PSA at each cycle, read by PCWG3 criteria, and considers it reliable 2 to 3 weeks after the second cycle, because of flare and delayed response (Kratochwil, 2023). PCWG3 advises continuing treatment despite early PSA rises in the first 12 weeks, if nothing else indicates progression (Scher, 2016). The two rules coexist: the second cycle usually falls near that mark, and the review on multimodal imaging describes flare of up to 12 weeks (Ranjan, 2026).

When PSA falls, the drop has prognostic value. In the VISION analysis, a PSA drop of 50 out of every 100 or more occurred in 46 out of every 100 men treated with lutetium-PSMA, versus 7.1 out of every 100 in the control group (Armstrong, 2024). An isolated rise in the first weeks, without clinical worsening and without new lesions, calls for caution before stopping treatment.

RECIP: PSMA PET as a ruler

RECIP 1.0 was created to standardize the reading of PSMA PET during lutetium-PSMA. The study included 124 men with PET at baseline and at an interim assessment around 12 weeks (Gafita, 2022). The classification combines the change in total volume of PSMA-positive disease with the appearance of new lesions, read by three nuclear medicine physicians blinded to outcome (Gafita, 2022). The authors started from an observation from practice: total disease burden can fall at the same time that new lesions appear (Gafita, 2022).

The categories separated prognoses well. Median overall survival was 8.3 months in progression, 13.1 months in stable disease and 21.7 months in partial response (Gafita, 2022). The combination of PSA and RECIP had a concordance index of 0.65, versus 0.62 for PSA alone, for identifying responders (Gafita, 2022). For identifying progression, the index was 0.66 for the combination, versus 0.63 with PSA alone (Gafita, 2022).

Visual reading also works. The review on multimodal imaging describes concordance of 0.89 between visual and quantitative RECIP, measured by kappa (Ranjan, 2026). This brings the criterion closer to services without segmentation software.

The joint guideline proposes restaging every 12 weeks, always with a second modality besides PSA (Kratochwil, 2023). The reasoning for using functional imaging to guide management is the same discussed in the text about what interim PET measures in lymphoma.

In neuroendocrine tumors

Well-differentiated neuroendocrine tumors grow slowly, and the response appears late. The guidance for peptide therapy places the first response assessment around 3 months after treatment and the following ones every 3 to 6 months (Bodei, 2013). The most important rule is to use the same imaging modality as the baseline (Bodei, 2013).

Switching exams along the way creates false signals. A lesion that looks new on PET may have always been there and gone unnoticed on a less sensitive scintigraphy. A lesion that looks smaller on MRI may be unchanged on CT. The response report must state which exam the comparison was made with.

Fourth step: follow toxicity

Follow-up is the longest part of the cycle and the easiest to lose. The person leaves the nuclear medicine service, returns to oncology and, months later, arrives at the emergency department or primary care with an abnormal blood count. Whoever sees them needs to know that they received a radioligand and what that can cause.

Marrow and kidney

Kidney and bone marrow are the critical organs of peptide therapy (Bodei, 2013). The guidance proposes a blood count every 2 to 4 weeks between cycles (Bodei, 2013). After treatment, the blood count continues every 8 to 12 weeks for 12 months and then twice a year (Bodei, 2013).

For lutetium-PSMA, the joint guideline calls for a blood count and creatinine 2 to 3 weeks after each cycle (Kratochwil, 2023). The same tests are repeated 4 to 6 weeks after the last cycle (Kratochwil, 2023).

In NETTER-1, severe hematologic toxicity was infrequent. Grade 3 or 4 neutropenia, thrombocytopenia and lymphopenia appeared in 1, 2 and 9 in every 100 patients, respectively (Strosberg, 2017). The study reported no signs of renal toxicity with a median follow-up of 14 months (Strosberg, 2017). That follow-up is short for renal effects, which tend to appear late.

Late myeloid neoplasms

The most feared complication is therapy-related myeloid neoplasm. In NETTER-1, one patient in the lutetium group had myelodysplastic syndrome, which corresponds to 0.9 in every 100 (Strosberg, 2017). In the final analysis of NETTER-1, with a median follow-up of 76.3 months, 2 of 111 patients in the lutetium group had myelodysplastic syndrome, and no new cases of myelodysplastic syndrome or acute myeloid leukemia appeared in the long follow-up (Strosberg, 2021).

These numbers contrast with those of a single tertiary center, which followed 173 patients treated with peptide therapy. Of these, 10 developed myeloid neoplasm, which corresponds to 5.8 in every 100, with a median of 23 months after treatment (Batista, 2026). Of the 10 patients, 8 died, with a median of 8 months after diagnosis (Batista, 2026). The authors cite a 2020 systematic review with a mean of 2.61 in every 100 (Batista, 2026).

In practice, persistent or new cytopenia months after the last cycle calls for a blood count with smear and hematologic evaluation, and not just an explanation by "treatment effect". The reasoning is the same described in the text about cancer after radiation exposure, with one difference: here the exposure was planned, and the follow-up can be planned too. The reading of the blood count after an exposure, with a focus on lymphocytes, appears in the text about differential diagnosis and exams in acute radiation syndrome.

Other effects that enter follow-up

Some effects depend on the target. PSMA is also expressed in the salivary glands, and the joint guideline includes the management of xerostomia in lutetium-PSMA care (Kratochwil, 2023). Dry mouth changes eating, sleep and dental health, and deserves a direct question at every return visit.

Pregnancy is an absolute contraindication to peptide therapy, and the guidance recommends avoiding pregnancy for 6 months after treatment (Bodei, 2013). With lutetium-PSMA, the joint guideline calls for contraception during treatment and for 3 months afterward, as a general precaution with radiopharmaceuticals (Kratochwil, 2023). The American label advises men with a partner of childbearing age to use effective contraception during treatment and for 14 weeks after the last dose (FDA, 2025). The 6-month window for peptide therapy therefore applies only to it. In people of childbearing age, the conversation about contraception is part of consent. Counseling after an exposure during pregnancy is the topic of the blog's Radiation series.

Building the cycle step by step

Back to the woman from the opening: the intense uptake in the liver metastases is the first piece. The team asks whether every lesion seen on CT takes up the analog, whether the histologic grade and the speed of progression fit well-differentiated disease, and whether there is a bulky liver lesion that justifies an FDG PET to look for discordant disease (Hindié, 2017).

Before treatment come a blood count, kidney function and the history of previous treatments. If there is carcinoid syndrome, coverage with a somatostatin analog goes into the plan (Bodei, 2013). Afterward, response is measured with the same scan as at baseline.

The table summarizes the cycle on the two platforms.

Step	Neuroendocrine tumors (SSTR)	Prostate cancer (PSMA)
Selection	Uptake in all target lesions; FDG if discordance is suspected	PSMA PET; FDG optional, useful in the liver
Before treating	Kidney, marrow, functional syndrome	Kidney, marrow, previous treatments
Infusion day	Amino acids, nausea, carcinoid crisis	Radiation protection precautions
Post-treatment imaging	Scintigraphy showing uptake	Scintigraphy or SPECT 1 to 2 days later
Response	First around 3 months, same modality	PSA after the flare and PET with RECIP
Follow-up	Blood count for years; late myeloid neoplasm	Blood count and creatinine; xerostomia

The data in the table come from the guidance for peptide therapy (Bodei, 2013) and for lutetium-PSMA (Kratochwil, 2023), with RECIP as described by Gafita (Gafita, 2022).

The team around the cycle

Each step involves more than one specialty. Oncology knows the disease and the previous treatments, nuclear medicine selects, treats and reads the response, radiology contributes the anatomy, nursing prepares the infusion and gives guidance for home, and hematology and nephrology join the follow-up. Some tasks often end up without an owner:

• recording in the chart that the person received a radioligand, with dates;
• agreeing on who orders and who reads the blood count after the last cycle;
• defining which scan will be the yardstick for response;
• telling the person about signs that call for contact, such as fever, bleeding and intense dry mouth;
• aligning with radiation oncology when local treatment is planned, as in the text on how to plan salvage;
• reviewing kidney function and marrow before each new line of treatment, as in the text on the limits of marrow and kidney.

How NEXO helps train this reasoning

In NEXO, the case of the day is fictional, with a real clinical presentation and up to six progressive clues, and you propose the diagnosis by looking for the code in ICD-10.

The limits on using the game are described on the NEXO economy page. If you research the teaching of clinical reasoning, the page for researchers explains how to get in touch. Professionals who write or review cases find information under curation and creators, and institutions can learn about the proposal on the partnerships page. Training differential diagnosis with clues and looking for the right code in ICD-10 complete this practice.

Frequently asked questions

What is theranostics?

It is the use of a molecule that shows the target on imaging and of a therapeutic version that delivers radiation to the same target, from radioactive iodine in the thyroid onward (Oldan, 2025).

Is it enough for the target to appear on imaging to indicate treatment?

Imaging of the target is necessary, but lesions without the target and with FDG uptake worsen the course and must enter the decision (Hindié, 2017).

Why can PSA rise at the start of lutetium-PSMA?

There may be a transient rise, the flare, and the guideline considers PSA reliable only 2 to 3 weeks after the second cycle (Kratochwil, 2023).

What is RECIP?

It is a response criterion for PSMA PET that combines the change in disease volume with new lesions and, added to PSA, predicts survival better (Gafita, 2022).

Which scan should be used to measure response in neuroendocrine tumors?

The same scan used at baseline, so that the comparison is valid (Bodei, 2013).

For how long should the blood count be followed?

For years. The guidance proposes a blood count for 12 months after treatment and then twice a year (Bodei, 2013).

Milestone inventory

• Origin: radioactive iodine, from the work of Saul Hertz, opens the idea of seeing and treating with the same substance (Oldan, 2025).
• MIBG: the first agent developed on purpose for imaging and treatment of adrenal medulla tumors and neuroblastoma (Oldan, 2025).
• 2013: the joint guidance for peptide therapy in neuroendocrine tumors organizes selection, care and follow-up (Bodei, 2013).
• 2015: studies of lutetium-PSMA begin to appear, and the PSMA-I&T ligand is developed (Zhang, 2021).
• 2017: NETTER-1 shows a gain in progression-free survival with lutetium-DOTATATE (Strosberg, 2017), and the NETPET score proposes reading FDG and receptor together (Hindié, 2017).
• 2021: VISION shows a survival gain with lutetium-PSMA-617 (Sartor, 2021).
• 2022: RECIP 1.0 standardizes response on PSMA PET (Gafita, 2022).
• 2023: the joint guideline for lutetium-PSMA defines selection, post-treatment imaging and the schedule of scans (Kratochwil, 2023).
• 2024: the PSA analysis in VISION shows the prognostic value of the early drop (Armstrong, 2024).
• 2025: a randomized trial tests dosimetry-based planning in peptide therapy (Kolodziej, 2025).
• 2026: a series with long follow-up describes late myeloid neoplasms (Batista, 2026), and a review organizes the role of each imaging scan in lutetium-PSMA (Ranjan, 2026).

Take-home messages

• Theranostics is a four-step cycle: select, treat, measure response and follow toxicity.
• Target imaging selects, but disease without a target and FDG positive can decide the course.
• PSA can rise at the start, and response is more reliable when it combines PSA and PET.
• In neuroendocrine tumors, response is measured with the same scan as at baseline.
• Kidney and marrow are the critical organs, and the blood count continues for years.
• Late cytopenia in someone who received a radioligand calls for hematological evaluation.

Want to train the reasoning that goes beyond the first positive scan? Play the case of the day on NEXO.

References

1. Armstrong AJ, Sartor O, de Bono J, et al. Association of Declining Prostate-specific Antigen Levels with Clinical Outcomes in Patients with Metastatic Castration-resistant Prostate Cancer Receiving [(177)Lu]Lu-PSMA-617 in the Phase 3 VISION Trial. Eur Urol. 2024. PMID 39242323. DOI 10.1016/j.eururo.2024.08.021.
2. Batista JPT, De La Iglesia M, Kim H, et al. Hematological neoplasms after PRRT: experience from long term follow-up of 173 patients in a tertiary center. Endocr Oncol. 2026. PMID 42694695. DOI 10.1530/EO-26-0038.
3. Bodei L, Mueller-Brand J, Baum RP, et al. The joint IAEA, EANM, and SNMMI practical guidance on peptide receptor radionuclide therapy (PRRNT) in neuroendocrine tumours. Eur J Nucl Med Mol Imaging. 2013. PMID 23389427. DOI 10.1007/s00259-012-2330-6.
4. FDA. Lutetium Lu 177 vipivotide tetraxetan injection: prescribing information. DailyMed. Label on DailyMed.
5. Gafita A, Rauscher I, Weber M, et al. Novel Framework for Treatment Response Evaluation Using PSMA PET/CT in Patients with Metastatic Castration-Resistant Prostate Cancer (RECIP 1.0): An International Multicenter Study. J Nucl Med. 2022. PMID 35422442. DOI 10.2967/jnumed.121.263072.
6. Hindié E. The NETPET Score: Combining FDG and Somatostatin Receptor Imaging for Optimal Management of Patients with Metastatic Well-Differentiated Neuroendocrine Tumors. Theranostics. 2017. PMID 28435455. DOI 10.7150/thno.19588.
7. Kolodziej M, Opalinska M, Mikolajczak R, et al. Dosimetry-guided peptide receptor radionuclide therapy in neuroendocrine tumors: interim safety analysis of the DUONEN trial. Front Endocrinol (Lausanne). 2025. PMID 41404513. DOI 10.3389/fendo.2025.1716247.
8. Kratochwil C, Fendler WP, Eiber M, et al. Joint EANM/SNMMI procedure guideline for the use of (177)Lu-labeled PSMA-targeted radioligand-therapy ((177)Lu-PSMA-RLT). Eur J Nucl Med Mol Imaging. 2023. PMID 37246997. DOI 10.1007/s00259-023-06255-8.
9. Oldan JD, Anugu A, Islam MZ, et al. Theranostics 2.0: Target-driven, artificial intelligence-enabled cancer therapy across tumor types. Diagn Interv Imaging. 2025. PMID 40752973. DOI 10.1016/j.diii.2025.07.003.
10. Ranjan A, Padden-Modi M, Abdel-Aty H, et al. The role of multimodality imaging in selection, response assessment, and follow-up of patients receiving (177)Lutetium-PSMA-therapy. Insights Imaging. 2026. PMID 41543663. DOI 10.1186/s13244-025-02151-x.
11. Sartor O, de Bono J, Chi KN, et al. Lutetium-177-PSMA-617 for Metastatic Castration-Resistant Prostate Cancer. N Engl J Med. 2021. PMID 34161051. DOI 10.1056/NEJMoa2107322.
12. Scher HI, Morris MJ, Stadler WM, et al. Trial Design and Objectives for Castration-Resistant Prostate Cancer: Updated Recommendations From the Prostate Cancer Clinical Trials Working Group 3. J Clin Oncol. 2016. PMID 26903579. DOI 10.1200/JCO.2015.64.2702.
13. Strosberg J, El-Haddad G, Wolin E, et al. Phase 3 Trial of (177)Lu-Dotatate for Midgut Neuroendocrine Tumors. N Engl J Med. 2017. PMID 28076709. DOI 10.1056/NEJMoa1607427.
14. Strosberg JR, Caplin ME, Kunz PL, et al. (177)Lu-Dotatate plus long-acting octreotide versus high‑dose long-acting octreotide in patients with midgut neuroendocrine tumours (NETTER-1): final overall survival and long-term safety results from an open-label, randomised, controlled, phase 3 trial. Lancet Oncol. 2021. PMID 34793718. DOI 10.1016/S1470-2045(21)00572-6.
15. Zhang H, Koumna S, Pouliot F, et al. PSMA Theranostics: Current Landscape and Future Outlook. Cancers (Basel). 2021. PMID 34439177. DOI 10.3390/cancers13164023.