1.

Introduction

Dirty necrosis was defined as the presence of cell detritus and inflammatory cells within the glandular lumina. Assessment for the presence or absence of dirty necrosis is often considered as a characteristic of colorectal carcinomas (CRCs). It mainly results from the breakdown of carcinoma cells, which are typically accumulated within the lumina of intact tubular glands, and is more likely related to spontaneous phenomena involving an insufficient vascular supply of the tumor.^1,2 CRCs arise through at least two different molecular genetic pathways in their carcinogenesis: MSI and chromosomal instability.³ The majority of CRCs arise through chromosomal instability, whereas approximately 15% of tumors arise in the setting of MSI caused by deficiencies in DNA mismatch repair (MMR) machinery. Investigation has shown that lack of dirty necrosis is a diagnostically useful feature of CRCs with DNA MSI.³ Statistical analysis demonstrated that according to the absence of dirty necrosis, sensitivity and specificity of histologic predictors of MSI status can reach 82.7% and 76.6%, respectively, and when combining the absence of dirty necrosis with tumor infiltrating lymphocytes (TILs) and any mucinous differentiation, the sensitivity can increase to about 100%.^3,4

As several studies have confirmed that MSI colorectal cancers have a better prognosis than microsatellite stable (MSS) tumors,^5–8 it will become necessary to identify all MSI CRCs so that different therapies are developed to exploit the genetic differences between MSI and MSS tumors. Although dirty necrosis is a simple feature that is readily identifiable by pathologists using routine hematoxylin and eosin (H&E)-stained sections, there are several disadvantages to performing histological biopsies, including sampling error, bleeding, crush artifacts, cost, and time-consuming pathological procedures.⁹ Nevertheless, these shortcomings will be surmounted if a multiphoton endoscope is used in vivo in the near future.

Multiphoton microscopy (MPM) relies on nonlinear optical processes such as second-harmonic generation (SHG) and two-photon excited fluorescence (TPEF) to achieve high resolution imaging of biological tissues and has been used for tissue imaging. This new imaging technique displays several advantages over these traditional imaging techniques,^10–16 such as real-time, being label-free, near-infrared excitation for superior optical penetration as well as lower photodamage, and it has the ability to detect the cellular and subcellular microstructures of tissues.^17–22 To date, we have no known previous reports that investigated the dirty necrosis in colorectal carcinoma using MPM and this is what motivated us to do this work.

2.

Materials and Methods

3.

Results and Discussion

Figure 1 shows representative TPEF/SHG images of normal rectal submucosa and the corresponding H&E light microscopic images. Normal rectal submucosa mainly consists of connective tissues including collagen and elastic fibers. Collagens, which have a morphology of fine mesh and are the predominant structural protein in rectum submucosa, are distinctively visualized in Fig. 1(a) via their strong SHG signal, while the elastic fibers, which present a morphology of long-ropes and are much sparser than collagen in rectal submucosa, are displayed very well in Fig. 1(b) by their TPEF signal.²⁵ Interestingly, in the overlay image of Fig. 1(c), the collagens turn yellowish because they have comparable SHG and TPEF signals.

Fig. 1

Representative two-photon excited fluorescence (TPEF)/ second-harmonic generation (SHG) images and corresponding hematoxylin and eosin (H&E) image of normal rectal submucosa. (a) SHG image of collagen (color-coded green). (b) TPEF image of collagen and elastic fibers (color-coded red). (c) Overlay of the SHG and TPEF images. (d) The corresponding H&E image ($40\times$). Scale bar: 100 μm.

In comparison to normal tissues, Fig. 2 displays representative SHG/TPEF images of cancerous rectal submucosa and the corresponding H&E image. The collagens as well as elastin fibers obviously display a significant loss compared to normal tissue. The residual collagens become very thin because many abnormal cells infiltrate and displace their locations by degrading collagen, whereas elastic fibers become thick and fragmented (blue arrows). An interest result of blood vessel was found in this layer. The blood vessel wall becomes distorted and uneven (yellow arrow) just because it is compressed by the proliferating cancer cells in cancerous tissue.^26,27

Fig. 2

Representative TPEF/SHG images and corresponding H&E image of cancerous rectal submucosa. (a) SHG image of collagen (color-coded green). (b) TPEF image of dirty necrosis, cancerous cells, blood vessel, collagen, and elastic fibers (color-coded red). (c) Overlay of the SHG and TPEF images. (d) The corresponding H&E image ($40\times$). Scale bar: 100 μm.

More importantly, there are many cancerous rectal glands with a disordered arrangement as well as an irregular shape, which are a typical characteristic of adenocarcinoma²⁸ and indicate that cancerous cells have invaded into the submucosa. Many abnormal cells with circular nuclei and irregular sizes can be clearly discerned in Fig. 2(b). They aggregate in this layer, resulting in high nuclear/cytoplasmic ratios and demonstrating the same morphological characteristics as those of the mucosa.^25,29 The most obvious feature is that there are many fluorophores from dirty necrosis which fill with the irregular glandular lumina (white arrows). This confirms the presence of dirty necrosis. By contrast, the MPM images of glandular lumina are dark because only mucus is in them (pink arrows). Thus, it is very easy to evaluate whether the dirty necrosis is absent or present based on the MPM. These qualitative morphological variations are very consistent with the paired histological sections as shown in Fig. 2(d). To the best of our knowledge, no other study of MPM imaging about dirty necrosis has been described.

In order to clearly present the microstructures of dirty necrosis and cancerous cells, Fig. 3 shows magnified SHG/TPEF images of the part of the white box in Fig. 2(c) and the corresponding H&E image. It is obvious that the signal intensity from dirty necrosis is far greater than that of cancerous cells. In detail, the mean and standard deviation of emission intensity from dirty necrosis is $1453.60\pm 459.62$, while from the cancerous cells is $329.02\pm 141.46$, that is, the intensity ratio of dirty necrosis over cancerous cells is approximately 4. Furthermore, to exactly show the characteristic and analyze the constituents of dirty necrosis, emission spectra of dirty necrosis and cancerous cells after subtraction of background were simultaneously obtained under the same conditions by using the image-guide spectral analysis method as shown in Fig. 4, which ensures that they can be compared on same scale. The TPEF signal with three peaks at around 475, 511, and 540 nm was detected. According to the previous study, the fluorescence peaks at 475 and 540 nm were responsible for nicotinamide adenine dinucleotide and flavin adenine dinucleotide, respectively, while the strong fluorescence peak at 511 nm corresponded to a cellular structural protein.^15,23,30

Fig. 3

Magnified image of the region of interest shown in Fig. 2(c) (white box), showing more detailed structural information of dirty necrosis and cancerous cells. (a) SHG image of collagen (color-coded green). (b) TPEF image of carcinomatous cells and dirty necrosis (color-coded red). (c) Overlay of the SHG and TPEF images. (d) The corresponding H&E image. Scale bar: 50 μm.

Fig. 4

Emission spectra of dirty necrosis within glandular lumina and cancerous cells from the cancerous rectal submucosa, obtained with an excitation wavelength of 810 nm.

Dirty necrosis is a constant and diagnostically important histologic feature differentiating types of colon cancers.^4,31,32 In particular, dirty necrosis is negatively associated with MSI and has become an indispensable factor index as one of the pathologic predictors of MSI in colorectal cancer.^2,3,33,34 MSI is the molecular fingerprint of a deficient MMR system and approximately 15% of CRCs arise in the setting of MSI.³⁵ Several studies have demonstrated that MSI CRCs have a better prognosis than MSS tumors. Hence, identification of MSI colorectal cancers is important, not only for the identification of hereditary nonpolyposis colorectal cancer syndrome but also for tailoring a successful therapy.^36,37

However, this is an expensive process that requires either DNA testing to identify MSI or immunoperoxidase staining to look for the loss of DNA MMR enzymes.^38–40 Recently, some researchers have described the morphological features of MSI tumors and find that the lack of dirty necrosis is significantly associated with MSI tumors.^3,4,31 This simple characteristic is easily identified from H&E stained histological section, but histological biopsies have several disadvantages including sampling error, bleeding, crush artifacts, and time-consuming pathological procedures.⁹ By contrast, MPM imaging of dirty necrosis can provide a real-time noninvasive optical diagnosis for differentiating an MSI colorectal carcinoma that should undergo evaluation to obtain better treatments.

Studies demonstrated that the lack of dirty necrosis was an indispensable factor in predicting MSI status and also an important factor in MMR defects that may cause MSI.^33,34,41 Greenson et al. showed that when it was used alone to predict MSI status, sensitivity and specificity can reach 82.7% and 76.6%, respectively. If combining the absence of dirty necrosis with TILs and any mucinous differentiation, the sensitivity can increase to about 100%.³ By contrast, if only dirty necrosis can be detected, the tumor may be excluded from MSI tumors. This will be helpful for differentiating MSI tumors from MSS tumors.

Our aim is to introduce a new approach for directly detecting dirty necrosis which may be helpful for MSI testing. MPM may be helpful to prescreen MSI tumors in vivo in the future. Although using DNA testing or immunoperoxidase techniques to identify MSI may be more precise for research purposes, it adds additional expense, which we think is not necessary to screen for MSI tumors.

4.

Conclusions

In summary, MPM was used to detect dirty necrosis in cancerous rectal submucosa. Our results suggested that MPM can be effective for identifying the dirty necrosis within glandular lumina in rectum cancer, except for differentiating the contents of rectal submucosa including collagen, elastic fibers, and blood vessel. These findings will be helpful for diagnosing MSI colorectal cancer in clinical care. With the continued advancement of miniature multiphoton-based endoscopy, it is foreseen that MPM can benefit in vivo diagnosis and treatment in the near future.