Many solid abdominal malignancies may cause peritoneal carcinomatosis. Tumor diffusion to the peritoneum represents a systemic cancer extension which, similar to the presence of hematogenous metastases, marks the substantial impossibility of definitively healing the patient. However, it is still possible to cure a selected subgroup of these patients, with results dependent either on biological aspects (for example, peritoneal seeding from ovarian is less aggressive than seeding from pancreatic and stomach tumors), either on stage in which the diagnosis is made . It is therefore very important to detect peritoneal carcinosis at an early stage, allowing to establish the prognosis with greater precision, giving to the patient the correct pathway of care. From the therapeutic point of view, the only possibilities of peritoneal carcinomatosis treatment are linked to a complete surgical reduction, eventually associated with intraperitoneal chemotherapy. Complete cytoreduction is possible only in the presence of small nodules . Unfortunately, the radiological instruments commonly used in the staging of abdominal tumors (CT, MRI, Pet, US) have a poor sensitivity for small peritoneal nodules [2, 3]. The best diagnostic tool is surgical exploration, most frequently by laparoscopic approach, associated with cytological examination on spontaneously present fluid or on peritoneal washing. However, even this technique has a limited sensitivity for peritoneal implants few millimeters-sized.
ICG, approved for clinical use by the Food and Drug Administration (FDA) since 1959, is the most commonly utilized fluorescent probe. It is a low-cost molecule, easy to use, widely available and with negligible toxicity . The use of ICG fluorescence in abdominal surgery has been introduced in recent years and represents a common tool for perfusion evaluation, extrahepatic bile duct anatomy, lymph node navigation and liver surgery [18–20]. ICG binds primarily to serum albumin and other serum globulins such as alpha1-lipoprotein, and then it circulates behaving like a macromolecule . In tumor tissues, such as peritoneal cancer implants, an “enhanced permeability and retention” (EPR) effect has been demonstrated, owing to tumor-induced angiogenesis, different metabolic activity and lack of efficient lymphatic drainage [22, 23]. ICG has theorical advantages as a possible contrast agent for macro- and microcirculatory tissue characterization, and consequently for EPR effect: it is uninfluenced by tissue optical properties and has half-life in plasma of few minutes [24, 25]. Some observations seem to indirectly confirm this point. In a series of patients with colorectal cancer, Filippello and Coll. reported that the fluorescence of carcinomatosis nodules was higher, and conversely that the rate of false-negative results was lower, in patients who did not receive bevacizumab compared with those who received the drug (76.3% and 65.0%, 42.9% and 53.8%, respectively). The anti-angiogenetic properties of bevacizumab may attenuate the enhanced permeability and retention of ICG . To date, it is not known whether these theoretical considerations have clinical confirmation. Furthermore the dosage, and above all the ideal timing of the injection are not clear. Looking at the results, the possible use of this technique compared to simple visual observation in terms of sensitivity, specificity and accuracy has not been clarified.
For this reasons, in the present systematic review we analyze methods and results of intravenous ICG injection in the diagnosis of peritoneal carcinosis. We found only 6 papers, all published in a short period of time (2015–2018) and by a few centers (Saint Etienne, Bruxelles, Naples, Tokyo, and Leiden). Modalities and timing of ICG injection were very similar. After some initial cases, almost all the Authors injected ICG at the time of anesthesia induction and detected fluorescence starting from 5’ after the injection, for a rather long period (someone up to 360'). Some experimental observations suggested that the best timing for ICG visualization due to the EPR is 6 hours after injection because owing to the rapid clearance of ICG, resulting in a better tumor-to-background ratio starting after 6 hours and lasting until 24 hours (20). In case of HCC (nodules in the peritoneum are usually not due to truly carcinosis, but rather implants due to the rupture of primitive cancer) the timing is different, as ICG is metabolized by normal hepatocytes and unexcreted because of bile ducts alteration. Indeed, papers reporting the usefulness of ICG in liver surgery always report intravenous injection from 1 to several days before surgery . In the case report of peritoneal implant of previously ruptured HCC excluded from the present systematic review, injection of 0,5 mg/kg ICG was performed 72 h before surgery . From a dosing point of view, virtually all series carry the same dosage (0.25 mg/kg), except one that carries twice and one that uses a fixed dose independently from weight (20 mg). All reported injected schedules are similar to that used in other areas for perfusion studies .
Overall, in the present systematic review we identified 71 patients. The vast majority (67 cases) were patients with known peritoneal carcinomatosis, undergoing elective surgery for cytoreduction and eventually HIPEC. This clinical setting was ideal for assessing the diagnostic performance of ICG on peritoneal carcinosis, being able to classify each lesion as ICG+/ICG -, malignant/benign by histological examination. By this way, 322 nodules were assessed. Statistical analysis confirmed that ICG is accurate, with sensitivity ranging from 72–100% and specificity from 54% and 100%. The average of these values was sensitivity 88.2% and specificity 77.8%. In some papers, subgroups of patients were also investigated. Colorectal carcinomas have been studied in relation to the mucinous component, concluding that this type of cancer has a poor affinity for ICG ; furthermore, in series investigating ovarian tumours, the accuracy of ICG on peritoneal scars after neoadjuvant chemotherapy was studied, reporting that this tissue also has little affinity for ICG . Some papers provided quantitative data, calculating the tumor-to-background ratio (TBR) [10, 13, 14]. TBR values around 2.0 have been consistently observed in ICG positive carcinosis nodules. However, in the remaining series, in which TBR has not been calculated, carcinosis nodules have still been detected as fluorescent. Aiming at an immediate and wider clinical spreading of this staging technique in clinical practice, we consider the quantitative aspect not essential.
Only in 4 cases peritoneal carcinosis was not known before surgery (all were open staging of ovarian (3 cases) and uterus cancer (1 case)) . Surprisingly, no paper focuses on ICG use during laparoscopic staging of abdominal cancers, with the exception of a single case report  and a series mainly focused on small hepatic surface metastases from periampullary cancers . Laparoscopic staging and therapeutic approach is now standard in colorectal cancers; for many other abdominal malignancies, such as gastric and HPB cancers, laparoscopic staging is widely performed [26, 27]. In laparoscopic staging, it would seem very easy to detect the fluorescence generated by ICG on the peritoneum, as many laparoscopic vision systems are currently equipped with fluorescence-driven surgery technology. Moreover, this setting appears even easier than the open one, in which a camera held by the hand, often quite cumbersome, may slowing down the intervention. For this reason, based on the promising data contained in this systematic review, we believe that the time has come to propose a multicenter prospective study to establish feasibility, technique and results of laparoscopic exploration of peritoneal carcinosis by fluorescence generated by ICG. A monocentric study on pancreatic malignancies has been proposed in 2019 . We personally believe the setting of gastric cancer may be also suitable for this kind of prospective study.
In the present systematic review we decided to consider only papers describing the use of ICG to detect nodules of peritoneal carcinosis, with the aim of evaluating the practical effectiveness of this technique in terms of an immediate translation in the daily practice. Some other fluorophores have been studied in humans, but none of them can currently be used outside experimental setting [29–32]. For the same reasons we decided not to consider studies on animal models and in vitro studies [20, 31–33]. Furthermore, we have not considered studies in which ICG is linked to molecules that bind directly to cancer cells [31, 34, 34]. This field of research, although very exciting from a theoretical point of view, as it could make fluorescence a truly oncological navigation, is however still far from a practical application .